diff options
author | Bernd Kuhls <bernd.kuhls@t-online.de> | 2015-03-23 18:34:30 +0100 |
---|---|---|
committer | Bernd Kuhls <bernd.kuhls@t-online.de> | 2015-03-23 18:34:30 +0100 |
commit | 8b0c8720e8ce08c9362e4eae570411d64c7e80a1 (patch) | |
tree | 83e9bc08534d9578e3dfe4fb24a3bddb642be118 /lib/enca/devel-docs/xml | |
parent | 264773be9ef0966ab2f060a2a7b17f057ea3c734 (diff) |
Remove lib/enca/
Second try to remove lib/enca/, let´s test whether
project/VS2010Express/XBMC for Windows.sln really needs libenca ;)
Forum thread for discussion
http://forum.kodi.tv/showthread.php?tid=221837
Diffstat (limited to 'lib/enca/devel-docs/xml')
-rw-r--r-- | lib/enca/devel-docs/xml/analyser.xml | 697 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/api-index-1.3..xml | 15 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/api-index-deprecated.xml | 5 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/api-index-full.xml | 130 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/auxiliary.xml | 177 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/encodings.xml | 286 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/enums.xml | 387 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/internal.xml | 1328 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/libenca-doc.bottom | 5 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/libenca-doc.top | 0 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/object_index.sgml | 11 | ||||
-rw-r--r-- | lib/enca/devel-docs/xml/tree_index.sgml | 5 |
12 files changed, 0 insertions, 3046 deletions
diff --git a/lib/enca/devel-docs/xml/analyser.xml b/lib/enca/devel-docs/xml/analyser.xml deleted file mode 100644 index 480d775bdc..0000000000 --- a/lib/enca/devel-docs/xml/analyser.xml +++ /dev/null @@ -1,697 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<refentry id="libenca-Analyser"> -<refmeta> -<refentrytitle role="top_of_page" id="libenca-Analyser.top_of_page">Analyser</refentrytitle> -<manvolnum>3</manvolnum> -<refmiscinfo>LIBENCA Library</refmiscinfo> -</refmeta> - -<refnamediv> -<refname>Analyser</refname> -<refpurpose>Basic analyser interface.</refpurpose> -</refnamediv> - -<refsynopsisdiv id="libenca-Analyser.synopsis" role="synopsis"> -<title role="synopsis.title">Synopsis</title> - -<synopsis> -<link linkend="EncaAnalyser">EncaAnalyser</link> <link linkend="enca-analyser-alloc">enca_analyser_alloc</link> (const <link linkend="char">char</link> *langname); -<link linkend="void">void</link> <link linkend="enca-analyser-free">enca_analyser_free</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="EncaEncoding">EncaEncoding</link> <link linkend="enca-analyse">enca_analyse</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size); -<link linkend="EncaEncoding">EncaEncoding</link> <link linkend="enca-analyse-const">enca_analyse_const</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size); -<link linkend="int">int</link> <link linkend="enca-errno">enca_errno</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -const <link linkend="char">char</link>* <link linkend="enca-strerror">enca_strerror</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> errnum); -<link linkend="int">int</link> <link linkend="enca-double-utf8-check">enca_double_utf8_check</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size); -<link linkend="int">int</link>* <link linkend="enca-double-utf8-get-candidates">enca_double_utf8_get_candidates</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="void">void</link> <link linkend="enca-set-multibyte">enca_set_multibyte</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> multibyte); -<link linkend="int">int</link> <link linkend="enca-get-multibyte">enca_get_multibyte</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="void">void</link> <link linkend="enca-set-interpreted-surfaces">enca_set_interpreted_surfaces</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> interpreted_surfaces); -<link linkend="int">int</link> <link linkend="enca-get-interpreted-surfaces">enca_get_interpreted_surfaces</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="void">void</link> <link linkend="enca-set-ambiguity">enca_set_ambiguity</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> ambiguity); -<link linkend="int">int</link> <link linkend="enca-get-ambiguity">enca_get_ambiguity</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="void">void</link> <link linkend="enca-set-filtering">enca_set_filtering</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> filtering); -<link linkend="int">int</link> <link linkend="enca-get-filtering">enca_get_filtering</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="void">void</link> <link linkend="enca-set-garbage-test">enca_set_garbage_test</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> garabage_test); -<link linkend="int">int</link> <link linkend="enca-get-garbage-test">enca_get_garbage_test</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="void">void</link> <link linkend="enca-set-termination-strictness">enca_set_termination_strictness</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> termination_strictness); -<link linkend="int">int</link> <link linkend="enca-get-termination-strictness">enca_get_termination_strictness</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="int">int</link> <link linkend="enca-set-significant">enca_set_significant</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="size-t">size_t</link> significant); -<link linkend="size-t">size_t</link> <link linkend="enca-get-significant">enca_get_significant</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -<link linkend="int">int</link> <link linkend="enca-set-threshold">enca_set_threshold</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="double">double</link> threshold); -<link linkend="double">double</link> <link linkend="enca-get-threshold">enca_get_threshold</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -</synopsis> -</refsynopsisdiv> - - - - - - - - - -<refsect1 id="libenca-Analyser.description" role="desc"> -<title role="desc.title">Description</title> -<para> -Basically you want to allocate an analyser with <link linkend="enca-analyser-alloc"><function>enca_analyser_alloc()</function></link> for some -language, use <link linkend="enca-analyse"><function>enca_analyse()</function></link> (or <link linkend="enca-analyse-const"><function>enca_analyse_const()</function></link>) on a buffer to find its -encoding, and interpret the results with something like <link linkend="enca-charset-name"><function>enca_charset_name()</function></link>. -The analyser then can be used for another buffer. Once you no longer need -it, call <link linkend="enca-analyser-free"><function>enca_analyser_free()</function></link> to release it. -</para> -<para> -A single working example is better than a hundred pages of reference manual. -</para> -<example> -<title>A minimal Enca library application – Czech encoding -detector.</title> -<programlisting> -#include <stdio.h> -#include <enca.h> -int -main(void) -{ - EncaAnalyser analyser; - EncaEncoding encoding; - unsigned char buffer[4096]; - size_t buflen; - buflen = fread(buffer, 1, 4096, stdin); - analyser = enca_analyser_alloc("cs"); - encoding = enca_analyse(analyser, buffer, buflen); - printf("Charset: %%s\n", enca_charset_name(encoding.charset, - ENCA_NAME_STYLE_HUMAN)); - enca_analyser_free(analyser); - return 0; -} -</programlisting> -</example> -<para> -The analyser has plenty of options, but generally you don't need to fiddle -with them, except <link linkend="enca-set-termination-strictness"><function>enca_set_termination_strictness()</function></link>. -</para> -<para> -All names prefixed with <function>ENCA_</function>, -<function>Enca</function>, <function>_Enca</function>, -or <function>enca_</function> should be treated as reserved and not used for -application function/variable/type/macro names. -</para> -</refsect1> - -<refsect1 id="libenca-Analyser.details" role="details"> -<title role="details.title">Details</title> -<refsect2 id="enca-analyser-alloc" role="function"> -<title>enca_analyser_alloc ()</title> -<indexterm zone="enca-analyser-alloc"><primary sortas="enca_analyser_alloc">enca_analyser_alloc</primary></indexterm><programlisting><link linkend="EncaAnalyser">EncaAnalyser</link> enca_analyser_alloc (const <link linkend="char">char</link> *langname);</programlisting> -<para> -Allocates an analyser and initializes it for language <parameter>language</parameter>. -</para> -<para> -The analyser, once crerated, can be used only for language for which it -was initialized. If you need to detect encodings of texts in more than one -language, you must allocate an analyser for each one. Note however, an -analyser may occupy a considerable amount of memory (a few hundreds of kB), -so it's generally not a good idea to have several hundreds of them floating -around. -</para> -<para> -<parameter>langname</parameter> is two-letter ISO 639:1989 language code. Locale names in form -language_territory and ISO-639 English language names also may be accepted -in the future. To be on the safe side, use only names returned by -<link linkend="enca-get-languages"><function>enca_get_languages()</function></link>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>langname</parameter> :</term> -<listitem><simpara> Language for which the analyser should be initialized. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The newly created <link linkend="EncaAnalyser"><type>EncaAnalyser</type></link> on success, <link linkend="NULL--CAPS"><type>NULL</type></link> on failure - (namely when <parameter>langname</parameter> is unknown or otherwise invalid). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-analyser-free" role="function"> -<title>enca_analyser_free ()</title> -<indexterm zone="enca-analyser-free"><primary sortas="enca_analyser_free">enca_analyser_free</primary></indexterm><programlisting><link linkend="void">void</link> enca_analyser_free (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Frees memory used by <link linkend="EncaAnalyser"><type>EncaAnalyser</type></link> <parameter>analyser</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser to be destroyed. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-analyse" role="function"> -<title>enca_analyse ()</title> -<indexterm zone="enca-analyse"><primary sortas="enca_analyse">enca_analyse</primary></indexterm><programlisting><link linkend="EncaEncoding">EncaEncoding</link> enca_analyse (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size);</programlisting> -<para> -Analyses <parameter>buffer</parameter> and finds its encoding. -</para> -<para> -The <parameter>buffer</parameter> is checked for 8bit encodings of language for which <parameter>analyser</parameter> -was initialized and for multibyte encodings, mostly independent on language -(unless disabled with <link linkend="enca-set-multibyte"><function>enca_set_multibyte()</function></link>). -</para> -<para> -The contents of <parameter>buffer</parameter> may be (and probably will be) modified during the -analyse. Use <link linkend="enca-analyse-const"><function>enca_analyse_const()</function></link> instead if this discomforts you.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser initialized for some language. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>buffer</parameter> :</term> -<listitem><simpara> Buffer to be analysed. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>size</parameter> :</term> -<listitem><simpara> Size of <parameter>buffer</parameter>. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Encoding of <parameter>buffer</parameter>. When charset part of return value is - <link linkend="ENCA-CS-UNKNOWN--CAPS"><type>ENCA_CS_UNKNOWN</type></link>, encoding was not determined. Check - <link linkend="enca-errno"><function>enca_errno()</function></link> for reason. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-analyse-const" role="function"> -<title>enca_analyse_const ()</title> -<indexterm zone="enca-analyse-const"><primary sortas="enca_analyse_const">enca_analyse_const</primary></indexterm><programlisting><link linkend="EncaEncoding">EncaEncoding</link> enca_analyse_const (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size);</programlisting> -<para> -Analyses <parameter>buffer</parameter> and finds its encoding. -</para> -<para> -The <parameter>buffer</parameter> is checked for 8bit encodings of language for which <parameter>analyser</parameter> -was initialized and for multibyte encodings, mostly independent on language -(unless disabled with <link linkend="enca-set-multibyte"><function>enca_set_multibyte()</function></link>). -</para> -<para> -This function never modifies <parameter>buffer</parameter> (can be even used with string literal -<parameter>buffer</parameter>) at the expense it's generally slower than <link linkend="enca-analyse"><function>enca_analyse()</function></link>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser initialized for some language. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>buffer</parameter> :</term> -<listitem><simpara> Buffer to be analysed. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>size</parameter> :</term> -<listitem><simpara> Size of <parameter>buffer</parameter>. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Encoding of <parameter>buffer</parameter>. When charset part of return value is - <link linkend="ENCA-CS-UNKNOWN--CAPS"><type>ENCA_CS_UNKNOWN</type></link>, encoding was not determined. Check - <link linkend="enca-errno"><function>enca_errno()</function></link> for reason. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-errno" role="function"> -<title>enca_errno ()</title> -<indexterm zone="enca-errno"><primary sortas="enca_errno">enca_errno</primary></indexterm><programlisting><link linkend="int">int</link> enca_errno (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns analyser error code. -</para> -<para> -The error code is not modified. However, any other analyser call i.e. -call to a function taking <parameter>analyser</parameter> as parameter can change the error code.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Error code of reason why last analyser call failed. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-strerror" role="function"> -<title>enca_strerror ()</title> -<indexterm zone="enca-strerror"><primary sortas="enca_strerror">enca_strerror</primary></indexterm><programlisting>const <link linkend="char">char</link>* enca_strerror (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> errnum);</programlisting> -<para> -Returns string describing the error code. -</para> -<para> -The returned string must be considered constant and must NOT be freed. -It is however gauranteed not to be modified on invalidated by subsequent -calls to any libenca functions, including <link linkend="enca-strerror"><function>enca_strerror()</function></link>. -</para> -<para> -The analyser error code is not changed for a successful call, and it set -to <link linkend="ENCA-EINVALUE--CAPS"><type>ENCA_EINVALUE</type></link> upon error.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>errnum</parameter> :</term> -<listitem><simpara> An analyser error code. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> String describing the error code. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-double-utf8-check" role="function"> -<title>enca_double_utf8_check ()</title> -<indexterm zone="enca-double-utf8-check"><primary sortas="enca_double_utf8_check">enca_double_utf8_check</primary></indexterm><programlisting><link linkend="int">int</link> enca_double_utf8_check (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size);</programlisting> -<para> -Checks buffer for double-UTF-8 encoding. -</para> -<para> -Double-UTF-8 encoding is the result of [errorneous] conversion of UTF-8 text -to UTF-8 again, as if it was in some 8bit charset. This is quite hard to -recover from. -</para> -<para> -The analayser mostly only determines what language will be assumed, -the rest of this test is independent on the main guessing routines. -When <parameter>buffer</parameter> doesn't containing UTF-8 text, the result is undefined -(namely, false positives are possible). -</para> -<para> -Calling this function when language is `none' has currently no effect.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state determinig the language for double-UTF-8 check. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>buffer</parameter> :</term> -<listitem><simpara> The buffer to be checked [<parameter>size</parameter>]. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>size</parameter> :</term> -<listitem><simpara> The size of <parameter>buffer</parameter>. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero, when <parameter>buffer</parameter> probably contains doubly-UTF-8 encoded text. - More precisely, it returns the number of charsets which are - possible candidates for source charset. You can then use - <link linkend="enca-double-utf8-get-candidates"><function>enca_double_utf8_get_candidates()</function></link> to retrieve the charsets. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-double-utf8-get-candidates" role="function"> -<title>enca_double_utf8_get_candidates ()</title> -<indexterm zone="enca-double-utf8-get-candidates"><primary sortas="enca_double_utf8_get_candidates">enca_double_utf8_get_candidates</primary></indexterm><programlisting><link linkend="int">int</link>* enca_double_utf8_get_candidates (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns array of double-UTF-8 source charset candidates from the last check. -</para> -<para> -The returned array should be freed by caller then no longer needed. Its -is the return value of the preceding <link linkend="enca-double-utf8-check"><function>enca_double_utf8_check()</function></link> call. -</para> -<para> -When called before any double-UTF-8 test has been performed yet or after -and unsuccessfull double-UTF-8 test, it returns NULL, but the result after -an unsuccessfull check should be considered undefined.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state for which double-UTF-8 candidates are to be - returned. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> An array containing charset id's of possible source charsets from - which the sample was doubly-UTF-8 encoded. The array may contain - only one value, but usually enca is not able to decide between - e.g. ISO-8859-2 and Win1250, thus more candidates are returned. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-set-multibyte" role="function"> -<title>enca_set_multibyte ()</title> -<indexterm zone="enca-set-multibyte"><primary sortas="enca_set_multibyte">enca_set_multibyte</primary></indexterm><programlisting><link linkend="void">void</link> enca_set_multibyte (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> multibyte);</programlisting> -<para> -Enables or disables multibyte encoding tests for <parameter>analyser</parameter>. -</para> -<para> -This option is enabled by default. -</para> -<para> -When multibyte encodings are disabled, only 8bit charsets are checked. -Disabling them for language with no 8bit charsets leaves only one thing -<link linkend="enca-analyse"><function>enca_analyse()</function></link> could test: whether the sample is purely 7bit ASCII or not -(the latter leading to analyser failure, of course). -</para> -<para> -Multibyte encoding detection is also affected by -<link linkend="enca-set-termination-strictness"><function>enca_set_termination_strictness()</function></link>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>multibyte</parameter> :</term> -<listitem><simpara> Whether multibyte encoding tests should be enabled (nonzero to - enable, zero to disable). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-multibyte" role="function" condition="since:1.3."> -<title>enca_get_multibyte ()</title> -<indexterm zone="enca-get-multibyte" role="1.3."><primary sortas="enca_get_multibyte">enca_get_multibyte</primary></indexterm><programlisting><link linkend="int">int</link> enca_get_multibyte (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns whether <parameter>analyser</parameter> can return multibyte encodings. -</para> -<para> -See <link linkend="enca-set-multibyte"><function>enca_set_multibyte()</function></link> for more detailed description of multibyte -encoding checking.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when multibyte encoding are possible, zero otherwise. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-interpreted-surfaces" role="function"> -<title>enca_set_interpreted_surfaces ()</title> -<indexterm zone="enca-set-interpreted-surfaces"><primary sortas="enca_set_interpreted_surfaces">enca_set_interpreted_surfaces</primary></indexterm><programlisting><link linkend="void">void</link> enca_set_interpreted_surfaces (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> interpreted_surfaces);</programlisting> -<para> -Enables or disables interpeted surfaces tests for <parameter>analyser</parameter>. -</para> -<para> -This option is enabled by default. -</para> -<para> -To allow simple applications which care about charset only and don't want -to wrestle with surfaces, neglecting surface information should not have -serious consequences. While ignoring EOL type surface is feasible, and -ignoring UCS byteorders may be acceptable in endian-homogenous environment; -ignoring the fact file is Quoted-Printable encoded can have disasterous -consequences. By disabling this option you can disable surfaces requiring -fundamental reinterpretation of the content, namely <link linkend="ENCA-SURFACE-QP--CAPS"><literal>ENCA_SURFACE_QP</literal></link> -and <link linkend="ENCA-SURFACE-EOL-BIN--CAPS"><literal>ENCA_SURFACE_EOL_BIN</literal></link> (thus probably making <link linkend="enca-analyse"><function>enca_analyse()</function></link> to fail on -such samples).</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>interpreted_surfaces</parameter> :</term> -<listitem><simpara> Whether interpreted surfaces tests should be enabled - (nonzero to allow, zero to disallow). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-interpreted-surfaces" role="function" condition="since:1.3."> -<title>enca_get_interpreted_surfaces ()</title> -<indexterm zone="enca-get-interpreted-surfaces" role="1.3."><primary sortas="enca_get_interpreted_surfaces">enca_get_interpreted_surfaces</primary></indexterm><programlisting><link linkend="int">int</link> enca_get_interpreted_surfaces (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns whether <parameter>analyser</parameter> allows interpreted surfaces. -</para> -<para> -See <link linkend="enca-set-interpreted-surfaces"><function>enca_set_interpreted_surfaces()</function></link> for more detailed description of -interpreted surfaces.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when interpreted surfaces are possible, zero otherwise. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-ambiguity" role="function"> -<title>enca_set_ambiguity ()</title> -<indexterm zone="enca-set-ambiguity"><primary sortas="enca_set_ambiguity">enca_set_ambiguity</primary></indexterm><programlisting><link linkend="void">void</link> enca_set_ambiguity (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> ambiguity);</programlisting> -<para> -Enables or disables ambiguous mode for <parameter>analyser</parameter>. -</para> -<para> -This option is disabled by default. -</para> -<para> -In ambiguous mode some result is returned even when the charset cannot be -determined uniquely, provided that sample contains only characters which -have the same meaning in all charsets under consideration.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>ambiguity</parameter> :</term> -<listitem><simpara> Whether result can be ambiguous (nonzero to allow, zero to - disallow). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-ambiguity" role="function" condition="since:1.3."> -<title>enca_get_ambiguity ()</title> -<indexterm zone="enca-get-ambiguity" role="1.3."><primary sortas="enca_get_ambiguity">enca_get_ambiguity</primary></indexterm><programlisting><link linkend="int">int</link> enca_get_ambiguity (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns whether <parameter>analyser</parameter> can return ambiguous results. -</para> -<para> -See <link linkend="enca-set-ambiguity"><function>enca_set_ambiguity()</function></link> for more detailed description of ambiguous results.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when ambiguous results are allowed, zero otherwise. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-filtering" role="function"> -<title>enca_set_filtering ()</title> -<indexterm zone="enca-set-filtering"><primary sortas="enca_set_filtering">enca_set_filtering</primary></indexterm><programlisting><link linkend="void">void</link> enca_set_filtering (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> filtering);</programlisting> -<para> -Enables or disables filters for <parameter>analyser</parameter>. -</para> -<para> -This option is enabled by default. -</para> -<para> -Various filters are used to filter out block of binary noise and box-drawing -characters that could otherwise confuse enca. In cases this is unwanted, -you can disable them by setting this option to zero.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>filtering</parameter> :</term> -<listitem><simpara> Whether filters should be enabled (nonzero to enable, zero to - disable). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-filtering" role="function" condition="since:1.3."> -<title>enca_get_filtering ()</title> -<indexterm zone="enca-get-filtering" role="1.3."><primary sortas="enca_get_filtering">enca_get_filtering</primary></indexterm><programlisting><link linkend="int">int</link> enca_get_filtering (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns whether <parameter>analyser</parameter> has filtering enabled. -</para> -<para> -See <link linkend="enca-set-filtering"><function>enca_set_filtering()</function></link> for more detailed description of filtering.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when filtering is enabled, zero otherwise. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-garbage-test" role="function"> -<title>enca_set_garbage_test ()</title> -<indexterm zone="enca-set-garbage-test"><primary sortas="enca_set_garbage_test">enca_set_garbage_test</primary></indexterm><programlisting><link linkend="void">void</link> enca_set_garbage_test (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> garabage_test);</programlisting> -<para> -Enables or disables garbage test for <parameter>analyser</parameter>. -</para> -<para> -This option is enabled by default. -</para> -<para> -To prevent white noise (and almost-white noise) from being accidentally -detected as some charset, a garbage test is used. In cases this is -unwanted, you can disable is by setting this option to zero.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>garabage_test</parameter> :</term> -<listitem><simpara> Whether garbage test should be allowed (nonzero to enable, - nzero to disable). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-garbage-test" role="function" condition="since:1.3."> -<title>enca_get_garbage_test ()</title> -<indexterm zone="enca-get-garbage-test" role="1.3."><primary sortas="enca_get_garbage_test">enca_get_garbage_test</primary></indexterm><programlisting><link linkend="int">int</link> enca_get_garbage_test (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns whether <parameter>analyser</parameter> has garbage test enabled. -</para> -<para> -See <link linkend="enca-set-garbage-test"><function>enca_set_garbage_test()</function></link> for more detailed description of garbage test.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when garbage test is enabled, zero otherwise. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-termination-strictness" role="function"> -<title>enca_set_termination_strictness ()</title> -<indexterm zone="enca-set-termination-strictness"><primary sortas="enca_set_termination_strictness">enca_set_termination_strictness</primary></indexterm><programlisting><link linkend="void">void</link> enca_set_termination_strictness (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="int">int</link> termination_strictness);</programlisting> -<para> -Enables or disables requiring multibyte sequences to be terminated correctly -at the end of sample. -</para> -<para> -This option is enabled by default. -</para> -<para> -The sample given to <link linkend="enca-analyse"><function>enca_analyse()</function></link> generally may not be a complete text -(e.g. for efficiency reasons). As a result, it may end in the middle of a -multibyte sequence. In this case, you should disable this option to -prevent rejecting some charset just because the sample don't terminate -correctly. On the other hand, when given sample contains whole text, you -should always enable this option to assure correctness of the result. -</para> -<para> -Note this option does NOT affect fixed character size encodings, like UCS-2 -and UCS-4; sample is never assumed to contain UCS-2 text when its size is -not even (and similarly for UCS-4).</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>termination_strictness</parameter> :</term> -<listitem><simpara> Whether multibyte sequences are required to be - terminated correctly at the end of sample - (nonzero to require, zero to relax). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-termination-strictness" role="function" condition="since:1.3."> -<title>enca_get_termination_strictness ()</title> -<indexterm zone="enca-get-termination-strictness" role="1.3."><primary sortas="enca_get_termination_strictness">enca_get_termination_strictness</primary></indexterm><programlisting><link linkend="int">int</link> enca_get_termination_strictness (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns whether <parameter>analyser</parameter> requires strict termination. -</para> -<para> -See <link linkend="enca-set-termination-strictness"><function>enca_set_termination_strictness()</function></link> for more detailed description of -termination strictness.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when strict termination is required, zero otherwise. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-significant" role="function"> -<title>enca_set_significant ()</title> -<indexterm zone="enca-set-significant"><primary sortas="enca_set_significant">enca_set_significant</primary></indexterm><programlisting><link linkend="int">int</link> enca_set_significant (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="size-t">size_t</link> significant);</programlisting> -<para> -Sets the minimal number of required significant characters. -</para> -<para> -The default value of this option is 10. -</para> -<para> -<link linkend="enca-analyse"><function>enca_analyse()</function></link> refuses to make a decision unles at least this number -of significant characters is found in sample. You may want to lower this -number for very short texts.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>significant</parameter> :</term> -<listitem><simpara> Minimal number of required significant characters. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Zero on success, nonzero on failure, i.e. when you passed zero - as <parameter>significant</parameter>. It sets analyser errno to ENCA_EINVALUE on - failure. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-significant" role="function" condition="since:1.3."> -<title>enca_get_significant ()</title> -<indexterm zone="enca-get-significant" role="1.3."><primary sortas="enca_get_significant">enca_get_significant</primary></indexterm><programlisting><link linkend="size-t">size_t</link> enca_get_significant (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns the minimum number of significant characters required by <parameter>analyser</parameter>. -</para> -<para> -See <link linkend="enca-set-significant"><function>enca_set_significant()</function></link> for more detailed description of required -significant characters.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The minimum number of significant characters. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> -<refsect2 id="enca-set-threshold" role="function"> -<title>enca_set_threshold ()</title> -<indexterm zone="enca-set-threshold"><primary sortas="enca_set_threshold">enca_set_threshold</primary></indexterm><programlisting><link linkend="int">int</link> enca_set_threshold (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser, - <link linkend="double">double</link> threshold);</programlisting> -<para> -Sets the minimal ratio between the most probable and the second most -probable charsets. -</para> -<para> -The default value of this option is 1.4142. -</para> -<para> -<link linkend="enca-analyse"><function>enca_analyse()</function></link> consideres the result known only when there's a clear gap -between the most probable and the second most probable charset -proababilities. Lower <parameter>threshold</parameter> values mean larger probability of a -mistake and smaller probability of not recognizing a charset; higher -<parameter>threshold</parameter> values the contrary. Threshold value of 2 is almost infinity.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>threshold</parameter> :</term> -<listitem><simpara> Minimal ratio between winner and second best guess. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Zero on success, nonzero on failure, i.e. when you passed value - smaller than 1.0 as <parameter>threshold</parameter>. It sets analyser errno to - ENCA_EINVALUE on failure. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-threshold" role="function" condition="since:1.3."> -<title>enca_get_threshold ()</title> -<indexterm zone="enca-get-threshold" role="1.3."><primary sortas="enca_get_threshold">enca_get_threshold</primary></indexterm><programlisting><link linkend="double">double</link> enca_get_threshold (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns the threshold value used by <parameter>analyser</parameter>. -</para> -<para> -See <link linkend="enca-set-threshold"><function>enca_set_threshold()</function></link> for more detailed threshold description.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The threshold value. - -</simpara></listitem></varlistentry> -</variablelist><para role="since">Since 1.3.</para></refsect2> - -</refsect1> - - - - -</refentry> diff --git a/lib/enca/devel-docs/xml/api-index-1.3..xml b/lib/enca/devel-docs/xml/api-index-1.3..xml deleted file mode 100644 index 9d73d9ff01..0000000000 --- a/lib/enca/devel-docs/xml/api-index-1.3..xml +++ /dev/null @@ -1,15 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE indexdiv PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<indexdiv> -<indexdiv><title>E</title> -<indexentry><primaryie linkends="enca-get-ambiguity"><link linkend="enca-get-ambiguity">enca_get_ambiguity</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-filtering"><link linkend="enca-get-filtering">enca_get_filtering</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-garbage-test"><link linkend="enca-get-garbage-test">enca_get_garbage_test</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-interpreted-surfaces"><link linkend="enca-get-interpreted-surfaces">enca_get_interpreted_surfaces</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-multibyte"><link linkend="enca-get-multibyte">enca_get_multibyte</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-significant"><link linkend="enca-get-significant">enca_get_significant</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-termination-strictness"><link linkend="enca-get-termination-strictness">enca_get_termination_strictness</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-threshold"><link linkend="enca-get-threshold">enca_get_threshold</link></primaryie></indexentry> -</indexdiv> -</indexdiv> diff --git a/lib/enca/devel-docs/xml/api-index-deprecated.xml b/lib/enca/devel-docs/xml/api-index-deprecated.xml deleted file mode 100644 index 113283526a..0000000000 --- a/lib/enca/devel-docs/xml/api-index-deprecated.xml +++ /dev/null @@ -1,5 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE indexdiv PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<indexdiv> -</indexdiv> diff --git a/lib/enca/devel-docs/xml/api-index-full.xml b/lib/enca/devel-docs/xml/api-index-full.xml deleted file mode 100644 index 1245d58abe..0000000000 --- a/lib/enca/devel-docs/xml/api-index-full.xml +++ /dev/null @@ -1,130 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE indexdiv PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<indexdiv> -<indexdiv><title>C</title> -<indexentry><primaryie linkends="CR--CAPS"><link linkend="CR--CAPS">CR</link></primaryie></indexentry> -</indexdiv> -<indexdiv><title>E</title> -<indexentry><primaryie linkends="ELEMENTS--CAPS"><link linkend="ELEMENTS--CAPS">ELEMENTS</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaAnalyserOptions"><link linkend="EncaAnalyserOptions">EncaAnalyserOptions</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaAnalyserState"><link linkend="EncaAnalyserState">EncaAnalyserState</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaCharsetFlags"><link linkend="EncaCharsetFlags">EncaCharsetFlags</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaCharsetInfo"><link linkend="EncaCharsetInfo">EncaCharsetInfo</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaEncoding"><link linkend="EncaEncoding">EncaEncoding</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaErrno"><link linkend="EncaErrno">EncaErrno</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaGuessFunc"><link linkend="EncaGuessFunc">EncaGuessFunc</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaHookFunc"><link linkend="EncaHookFunc">EncaHookFunc</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaLanguageHookData1CS"><link linkend="EncaLanguageHookData1CS">EncaLanguageHookData1CS</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaLanguageHookDataEOL"><link linkend="EncaLanguageHookDataEOL">EncaLanguageHookDataEOL</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaLanguageInfo"><link linkend="EncaLanguageInfo">EncaLanguageInfo</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaNameStyle"><link linkend="EncaNameStyle">EncaNameStyle</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaSurface"><link linkend="EncaSurface">EncaSurface</link></primaryie></indexentry> -<indexentry><primaryie linkends="EncaUTFCheckData"><link linkend="EncaUTFCheckData">EncaUTFCheckData</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-analyse"><link linkend="enca-analyse">enca_analyse</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-analyser-alloc"><link linkend="enca-analyser-alloc">enca_analyser_alloc</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-analyser-free"><link linkend="enca-analyser-free">enca_analyser_free</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-analyser-language"><link linkend="enca-analyser-language">enca_analyser_language</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-analyse-const"><link linkend="enca-analyse-const">enca_analyse_const</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charsets-subset-identical"><link linkend="enca-charsets-subset-identical">enca_charsets_subset_identical</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-has-ucs2-map"><link linkend="enca-charset-has-ucs2-map">enca_charset_has_ucs2_map</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-16bit"><link linkend="enca-charset-is-16bit">enca_charset_is_16bit</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-32bit"><link linkend="enca-charset-is-32bit">enca_charset_is_32bit</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-7bit"><link linkend="enca-charset-is-7bit">enca_charset_is_7bit</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-8bit"><link linkend="enca-charset-is-8bit">enca_charset_is_8bit</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-binary"><link linkend="enca-charset-is-binary">enca_charset_is_binary</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-fixed"><link linkend="enca-charset-is-fixed">enca_charset_is_fixed</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-known"><link linkend="enca-charset-is-known">enca_charset_is_known</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-multibyte"><link linkend="enca-charset-is-multibyte">enca_charset_is_multibyte</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-regular"><link linkend="enca-charset-is-regular">enca_charset_is_regular</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-is-variable"><link linkend="enca-charset-is-variable">enca_charset_is_variable</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-name"><link linkend="enca-charset-name">enca_charset_name</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-natural-surface"><link linkend="enca-charset-natural-surface">enca_charset_natural_surface</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-properties"><link linkend="enca-charset-properties">enca_charset_properties</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-charset-ucs2-map"><link linkend="enca-charset-ucs2-map">enca_charset_ucs2_map</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-csname"><link linkend="enca-csname">enca_csname</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-CS-UNKNOWN--CAPS"><link linkend="ENCA-CS-UNKNOWN--CAPS">ENCA_CS_UNKNOWN</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-double-utf8-check"><link linkend="enca-double-utf8-check">enca_double_utf8_check</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-double-utf8-destroy"><link linkend="enca-double-utf8-destroy">enca_double_utf8_destroy</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-double-utf8-get-candidates"><link linkend="enca-double-utf8-get-candidates">enca_double_utf8_get_candidates</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-double-utf8-init"><link linkend="enca-double-utf8-init">enca_double_utf8_init</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-eol-surface"><link linkend="enca-eol-surface">enca_eol_surface</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-errno"><link linkend="enca-errno">enca_errno</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-filter-boxdraw"><link linkend="enca-filter-boxdraw">enca_filter_boxdraw</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-find-max-sec"><link linkend="enca-find-max-sec">enca_find_max_sec</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-free"><link linkend="enca-free">enca_free</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-ambiguity"><link linkend="enca-get-ambiguity">enca_get_ambiguity</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-charset-aliases"><link linkend="enca-get-charset-aliases">enca_get_charset_aliases</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-charset-similarity-matrix"><link linkend="enca-get-charset-similarity-matrix">enca_get_charset_similarity_matrix</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-filtering"><link linkend="enca-get-filtering">enca_get_filtering</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-garbage-test"><link linkend="enca-get-garbage-test">enca_get_garbage_test</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-interpreted-surfaces"><link linkend="enca-get-interpreted-surfaces">enca_get_interpreted_surfaces</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-languages"><link linkend="enca-get-languages">enca_get_languages</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-language-charsets"><link linkend="enca-get-language-charsets">enca_get_language_charsets</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-multibyte"><link linkend="enca-get-multibyte">enca_get_multibyte</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-significant"><link linkend="enca-get-significant">enca_get_significant</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-surface-name"><link linkend="enca-get-surface-name">enca_get_surface_name</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-termination-strictness"><link linkend="enca-get-termination-strictness">enca_get_termination_strictness</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-get-threshold"><link linkend="enca-get-threshold">enca_get_threshold</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-guess-destroy"><link linkend="enca-guess-destroy">enca_guess_destroy</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-guess-init"><link linkend="enca-guess-init">enca_guess_init</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-BE--CAPS"><link linkend="ENCA-LANGUAGE-BE--CAPS">ENCA_LANGUAGE_BE</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-BG--CAPS"><link linkend="ENCA-LANGUAGE-BG--CAPS">ENCA_LANGUAGE_BG</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-CS--CAPS"><link linkend="ENCA-LANGUAGE-CS--CAPS">ENCA_LANGUAGE_CS</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-language-destroy"><link linkend="enca-language-destroy">enca_language_destroy</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-language-english-name"><link linkend="enca-language-english-name">enca_language_english_name</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-ET--CAPS"><link linkend="ENCA-LANGUAGE-ET--CAPS">ENCA_LANGUAGE_ET</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-language-hook-eol"><link linkend="enca-language-hook-eol">enca_language_hook_eol</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-language-hook-ncs"><link linkend="enca-language-hook-ncs">enca_language_hook_ncs</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-HR--CAPS"><link linkend="ENCA-LANGUAGE-HR--CAPS">ENCA_LANGUAGE_HR</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-HU--CAPS"><link linkend="ENCA-LANGUAGE-HU--CAPS">ENCA_LANGUAGE_HU</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-language-init"><link linkend="enca-language-init">enca_language_init</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-LT--CAPS"><link linkend="ENCA-LANGUAGE-LT--CAPS">ENCA_LANGUAGE_LT</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-LV--CAPS"><link linkend="ENCA-LANGUAGE-LV--CAPS">ENCA_LANGUAGE_LV</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-PL--CAPS"><link linkend="ENCA-LANGUAGE-PL--CAPS">ENCA_LANGUAGE_PL</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-RU--CAPS"><link linkend="ENCA-LANGUAGE-RU--CAPS">ENCA_LANGUAGE_RU</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-SK--CAPS"><link linkend="ENCA-LANGUAGE-SK--CAPS">ENCA_LANGUAGE_SK</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-SL--CAPS"><link linkend="ENCA-LANGUAGE-SL--CAPS">ENCA_LANGUAGE_SL</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-LANGUAGE-UK--CAPS"><link linkend="ENCA-LANGUAGE-UK--CAPS">ENCA_LANGUAGE_UK</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-malloc"><link linkend="enca-malloc">enca_malloc</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-name-to-charset"><link linkend="enca-name-to-charset">enca_name_to_charset</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-name-to-surface"><link linkend="enca-name-to-surface">enca_name_to_surface</link></primaryie></indexentry> -<indexentry><primaryie linkends="ENCA-NOT-A-CHAR--CAPS"><link linkend="ENCA-NOT-A-CHAR--CAPS">ENCA_NOT_A_CHAR</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-number-of-charsets"><link linkend="enca-number-of-charsets">enca_number_of_charsets</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-pair-analyse"><link linkend="enca-pair-analyse">enca_pair_analyse</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-pair-destroy"><link linkend="enca-pair-destroy">enca_pair_destroy</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-pair-init"><link linkend="enca-pair-init">enca_pair_init</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-parse-encoding-name"><link linkend="enca-parse-encoding-name">enca_parse_encoding_name</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-realloc"><link linkend="enca-realloc">enca_realloc</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-ambiguity"><link linkend="enca-set-ambiguity">enca_set_ambiguity</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-filtering"><link linkend="enca-set-filtering">enca_set_filtering</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-garbage-test"><link linkend="enca-set-garbage-test">enca_set_garbage_test</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-interpreted-surfaces"><link linkend="enca-set-interpreted-surfaces">enca_set_interpreted_surfaces</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-multibyte"><link linkend="enca-set-multibyte">enca_set_multibyte</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-significant"><link linkend="enca-set-significant">enca_set_significant</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-termination-strictness"><link linkend="enca-set-termination-strictness">enca_set_termination_strictness</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-set-threshold"><link linkend="enca-set-threshold">enca_set_threshold</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-stpcpy"><link linkend="enca-stpcpy">enca_stpcpy</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-strappend"><link linkend="enca-strappend">enca_strappend</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-strconcat"><link linkend="enca-strconcat">enca_strconcat</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-strdup"><link linkend="enca-strdup">enca_strdup</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-strerror"><link linkend="enca-strerror">enca_strerror</link></primaryie></indexentry> -<indexentry><primaryie linkends="enca-strstr"><link linkend="enca-strstr">enca_strstr</link></primaryie></indexentry> -<indexentry><primaryie linkends="EPSILON--CAPS"><link linkend="EPSILON--CAPS">EPSILON</link></primaryie></indexentry> -</indexdiv> -<indexdiv><title>F</title> -<indexentry><primaryie linkends="FILL-NONLETTER--CAPS"><link linkend="FILL-NONLETTER--CAPS">FILL_NONLETTER</link></primaryie></indexentry> -</indexdiv> -<indexdiv><title>L</title> -<indexentry><primaryie linkends="LF--CAPS"><link linkend="LF--CAPS">LF</link></primaryie></indexentry> -</indexdiv> -<indexdiv><title>M</title> -<indexentry><primaryie linkends="MAKE-HOOK-LINE--CAPS"><link linkend="MAKE-HOOK-LINE--CAPS">MAKE_HOOK_LINE</link></primaryie></indexentry> -</indexdiv> -<indexdiv><title>N</title> -<indexentry><primaryie linkends="NEW--CAPS"><link linkend="NEW--CAPS">NEW</link></primaryie></indexentry> -</indexdiv> -<indexdiv><title>R</title> -<indexentry><primaryie linkends="RENEW--CAPS"><link linkend="RENEW--CAPS">RENEW</link></primaryie></indexentry> -</indexdiv> -</indexdiv> diff --git a/lib/enca/devel-docs/xml/auxiliary.xml b/lib/enca/devel-docs/xml/auxiliary.xml deleted file mode 100644 index 6008844e0f..0000000000 --- a/lib/enca/devel-docs/xml/auxiliary.xml +++ /dev/null @@ -1,177 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<refentry id="libenca-Auxiliary-Functions"> -<refmeta> -<refentrytitle role="top_of_page" id="libenca-Auxiliary-Functions.top_of_page">Auxiliary functions</refentrytitle> -<manvolnum>3</manvolnum> -<refmiscinfo>LIBENCA Library</refmiscinfo> -</refmeta> - -<refnamediv> -<refname>Auxiliary functions</refname> -<refpurpose>Variouis auxiliary functions and informations about libenca.</refpurpose> -</refnamediv> - -<refsynopsisdiv id="libenca-Auxiliary-Functions.synopsis" role="synopsis"> -<title role="synopsis.title">Synopsis</title> - -<synopsis> -<link linkend="int">int</link> <link linkend="enca-charset-has-ucs2-map">enca_charset_has_ucs2_map</link> (<link linkend="int">int</link> charset); -<link linkend="int">int</link> <link linkend="enca-charset-ucs2-map">enca_charset_ucs2_map</link> (<link linkend="int">int</link> charset, - unsigned <link linkend="int">int</link> *buffer); -<link linkend="size-t">size_t</link> <link linkend="enca-number-of-charsets">enca_number_of_charsets</link> (void); -const <link linkend="char">char</link>* <link linkend="enca-analyser-language">enca_analyser_language</link> (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser); -const <link linkend="char">char</link>** <link linkend="enca-get-languages">enca_get_languages</link> (<link linkend="size-t">size_t</link> *n); -const <link linkend="char">char</link>* <link linkend="enca-language-english-name">enca_language_english_name</link> (const <link linkend="char">char</link> *lang); -<link linkend="int">int</link>* <link linkend="enca-get-language-charsets">enca_get_language_charsets</link> (const <link linkend="char">char</link> *langname, - <link linkend="size-t">size_t</link> *n); -</synopsis> -</refsynopsisdiv> - - - - - - - - - -<refsect1 id="libenca-Auxiliary-Functions.description" role="desc"> -<title role="desc.title">Description</title> -<para> -</para> -</refsect1> - -<refsect1 id="libenca-Auxiliary-Functions.details" role="details"> -<title role="details.title">Details</title> -<refsect2 id="enca-charset-has-ucs2-map" role="function"> -<title>enca_charset_has_ucs2_map ()</title> -<indexterm zone="enca-charset-has-ucs2-map"><primary sortas="enca_charset_has_ucs2_map">enca_charset_has_ucs2_map</primary></indexterm><programlisting><link linkend="int">int</link> enca_charset_has_ucs2_map (<link linkend="int">int</link> charset);</programlisting> -<para> -Find whether we have charset -> UCS-2 map.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset</parameter> :</term> -<listitem><simpara> An 8bit charset. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when the map is available, zero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-ucs2-map" role="function"> -<title>enca_charset_ucs2_map ()</title> -<indexterm zone="enca-charset-ucs2-map"><primary sortas="enca_charset_ucs2_map">enca_charset_ucs2_map</primary></indexterm><programlisting><link linkend="int">int</link> enca_charset_ucs2_map (<link linkend="int">int</link> charset, - unsigned <link linkend="int">int</link> *buffer);</programlisting> -<para> -Creates map from an 8bit charset to UCS-2. -</para> -<para> -The buffer must be at least 256*sizeof(unsigned int) long to hold the map.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset</parameter> :</term> -<listitem><simpara> An 8bit charset. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>buffer</parameter> :</term> -<listitem><simpara> Buffer to store the map to. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when the map was successfully created, zero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-number-of-charsets" role="function"> -<title>enca_number_of_charsets ()</title> -<indexterm zone="enca-number-of-charsets"><primary sortas="enca_number_of_charsets">enca_number_of_charsets</primary></indexterm><programlisting><link linkend="size-t">size_t</link> enca_number_of_charsets (void);</programlisting> -<para> -Returns number of known charsets. -</para> -<para> -Charsets idetifiers are assigned successively starting from zero, so last -charset has identifier <link linkend="enca-number-of-charsets"><function>enca_number_of_charsets()</function></link> - 1.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The number of charsets. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-analyser-language" role="function"> -<title>enca_analyser_language ()</title> -<indexterm zone="enca-analyser-language"><primary sortas="enca_analyser_language">enca_analyser_language</primary></indexterm><programlisting>const <link linkend="char">char</link>* enca_analyser_language (<link linkend="EncaAnalyser">EncaAnalyser</link> analyser);</programlisting> -<para> -Returns name of language which was <parameter>analyser</parameter> initialized for. -</para> -<para> -The returned string must be considered constant and must NOT be freed.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The language name. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-languages" role="function"> -<title>enca_get_languages ()</title> -<indexterm zone="enca-get-languages"><primary sortas="enca_get_languages">enca_get_languages</primary></indexterm><programlisting>const <link linkend="char">char</link>** enca_get_languages (<link linkend="size-t">size_t</link> *n);</programlisting> -<para> -Returns list of known languages. -</para> -<para> -The returned strings are two-letter ISO-639 language codes, the same as -<link linkend="enca-analyser-alloc"><function>enca_analyser_alloc()</function></link> accepts. -</para> -<para> -The list of languages has to be freed by caller; the strings themselves -must be considered constant and must NOT be freed.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>n</parameter> :</term> -<listitem><simpara> The number of languages will be stored here. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The list of languages, storing their number into *<parameter>n</parameter>. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-language-english-name" role="function"> -<title>enca_language_english_name ()</title> -<indexterm zone="enca-language-english-name"><primary sortas="enca_language_english_name">enca_language_english_name</primary></indexterm><programlisting>const <link linkend="char">char</link>* enca_language_english_name (const <link linkend="char">char</link> *lang);</programlisting> -<para> -Returns an English name of a language given its ISO-639 code. -</para> -<para> -The returned string must be considered constant and must NOT be freed.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>lang</parameter> :</term> -<listitem><simpara> A two-letter language code, such as obtained from - <link linkend="enca-analyser-language"><function>enca_analyser_language()</function></link> or <link linkend="enca-get-languages"><function>enca_get_languages()</function></link>. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The English language name. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-language-charsets" role="function"> -<title>enca_get_language_charsets ()</title> -<indexterm zone="enca-get-language-charsets"><primary sortas="enca_get_language_charsets">enca_get_language_charsets</primary></indexterm><programlisting><link linkend="int">int</link>* enca_get_language_charsets (const <link linkend="char">char</link> *langname, - <link linkend="size-t">size_t</link> *n);</programlisting> -<para> -Returns list of identifiers of charsets supported for language <parameter>language</parameter>. -</para> -<para> -The list of charset identifiers has to be freed by caller.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>langname</parameter> :</term> -<listitem><simpara> Two-letter ISO-639 language code. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>n</parameter> :</term> -<listitem><simpara> The number of charsets will be stored here. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The list of charsets, storing their number into *<parameter>n</parameter>. When language - contains no charsets or <parameter>langname</parameter> is invalid, <link linkend="NULL--CAPS"><type>NULL</type></link> is returned - and zero stored into *<parameter>n</parameter>. -</simpara></listitem></varlistentry> -</variablelist></refsect2> - -</refsect1> - - - - -</refentry> diff --git a/lib/enca/devel-docs/xml/encodings.xml b/lib/enca/devel-docs/xml/encodings.xml deleted file mode 100644 index 761e5ed660..0000000000 --- a/lib/enca/devel-docs/xml/encodings.xml +++ /dev/null @@ -1,286 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<refentry id="libenca-Charsets-and-Surfaces"> -<refmeta> -<refentrytitle role="top_of_page" id="libenca-Charsets-and-Surfaces.top_of_page">Charsets and Surfaces</refentrytitle> -<manvolnum>3</manvolnum> -<refmiscinfo>LIBENCA Library</refmiscinfo> -</refmeta> - -<refnamediv> -<refname>Charsets and Surfaces</refname> -<refpurpose>Functions and macros for getting something sensible from <link linkend="EncaEncoding"><type>EncaEncoding</type></link>.</refpurpose> -</refnamediv> - -<refsynopsisdiv id="libenca-Charsets-and-Surfaces.synopsis" role="synopsis"> -<title role="synopsis.title">Synopsis</title> - -<synopsis> -const <link linkend="char">char</link>* <link linkend="enca-charset-name">enca_charset_name</link> (<link linkend="int">int</link> charset, - <link linkend="EncaNameStyle">EncaNameStyle</link> whatname); -const <link linkend="char">char</link>** <link linkend="enca-get-charset-aliases">enca_get_charset_aliases</link> (<link linkend="int">int</link> charset, - <link linkend="size-t">size_t</link> *n); -<link linkend="char">char</link>* <link linkend="enca-get-surface-name">enca_get_surface_name</link> (<link linkend="EncaSurface">EncaSurface</link> surface, - <link linkend="EncaNameStyle">EncaNameStyle</link> whatname); -<link linkend="EncaEncoding">EncaEncoding</link> <link linkend="enca-parse-encoding-name">enca_parse_encoding_name</link> (const <link linkend="char">char</link> *name); -<link linkend="EncaSurface">EncaSurface</link> <link linkend="enca-charset-natural-surface">enca_charset_natural_surface</link> (<link linkend="int">int</link> charset); -<link linkend="EncaCharsetFlags">EncaCharsetFlags</link> <link linkend="enca-charset-properties">enca_charset_properties</link> (<link linkend="int">int</link> charset); -#define <link linkend="enca-charset-is-7bit">enca_charset_is_7bit</link> (cs) -#define <link linkend="enca-charset-is-8bit">enca_charset_is_8bit</link> (cs) -#define <link linkend="enca-charset-is-16bit">enca_charset_is_16bit</link> (cs) -#define <link linkend="enca-charset-is-32bit">enca_charset_is_32bit</link> (cs) -#define <link linkend="enca-charset-is-fixed">enca_charset_is_fixed</link> (cs) -#define <link linkend="enca-charset-is-variable">enca_charset_is_variable</link> (cs) -#define <link linkend="enca-charset-is-binary">enca_charset_is_binary</link> (cs) -#define <link linkend="enca-charset-is-regular">enca_charset_is_regular</link> (cs) -#define <link linkend="enca-charset-is-multibyte">enca_charset_is_multibyte</link> (cs) -#define <link linkend="enca-charset-is-known">enca_charset_is_known</link> (cs) -</synopsis> -</refsynopsisdiv> - - - - - - - - - -<refsect1 id="libenca-Charsets-and-Surfaces.description" role="desc"> -<title role="desc.title">Description</title> -<para> -</para> -</refsect1> - -<refsect1 id="libenca-Charsets-and-Surfaces.details" role="details"> -<title role="details.title">Details</title> -<refsect2 id="enca-charset-name" role="function"> -<title>enca_charset_name ()</title> -<indexterm zone="enca-charset-name"><primary sortas="enca_charset_name">enca_charset_name</primary></indexterm><programlisting>const <link linkend="char">char</link>* enca_charset_name (<link linkend="int">int</link> charset, - <link linkend="EncaNameStyle">EncaNameStyle</link> whatname);</programlisting> -<para> -Translates numeric charset id <parameter>charset</parameter> to some kind of name.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset</parameter> :</term> -<listitem><simpara> A charset id. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>whatname</parameter> :</term> -<listitem><simpara> Teh type of name you request. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The requested charset name; <link linkend="NULL--CAPS"><type>NULL</type></link> for invalid <parameter>whatname</parameter> or -<parameter>charset</parameter>, or when <parameter>whatname</parameter> name doesn't exist for charset <parameter>charset</parameter> -(<link linkend="ENCA-CS-UNKNOWN--CAPS"><type>ENCA_CS_UNKNOWN</type></link> is OK). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-charset-aliases" role="function"> -<title>enca_get_charset_aliases ()</title> -<indexterm zone="enca-get-charset-aliases"><primary sortas="enca_get_charset_aliases">enca_get_charset_aliases</primary></indexterm><programlisting>const <link linkend="char">char</link>** enca_get_charset_aliases (<link linkend="int">int</link> charset, - <link linkend="size-t">size_t</link> *n);</programlisting> -<para> -Returns list of accepted aliases for charset <parameter>charset</parameter>. -</para> -<para> -The list of aliases has to be freed by caller; the strings themselves -must be considered constant and must NOT be freed.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset</parameter> :</term> -<listitem><simpara> A charset id. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>n</parameter> :</term> -<listitem><simpara> The number of aliases will be stored here. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The list of aliases, storing their number into *<parameter>n</parameter>; <link linkend="NULL--CAPS"><type>NULL</type></link> for -invalid <parameter>charset</parameter> (*<parameter>n</parameter> is zero then). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-surface-name" role="function"> -<title>enca_get_surface_name ()</title> -<indexterm zone="enca-get-surface-name"><primary sortas="enca_get_surface_name">enca_get_surface_name</primary></indexterm><programlisting><link linkend="char">char</link>* enca_get_surface_name (<link linkend="EncaSurface">EncaSurface</link> surface, - <link linkend="EncaNameStyle">EncaNameStyle</link> whatname);</programlisting> -<para> -Constructs surface name from surface flags <parameter>surface</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>surface</parameter> :</term> -<listitem><simpara> A surface. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>whatname</parameter> :</term> -<listitem><simpara> The type of name you request. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The requested surface name; <link linkend="NULL--CAPS"><type>NULL</type></link> for invalid <parameter>whatname</parameter>; empty -string for naming style not supporting surfaces. In all cases, the -returned string must be freed by caller when no longer used. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-parse-encoding-name" role="function"> -<title>enca_parse_encoding_name ()</title> -<indexterm zone="enca-parse-encoding-name"><primary sortas="enca_parse_encoding_name">enca_parse_encoding_name</primary></indexterm><programlisting><link linkend="EncaEncoding">EncaEncoding</link> enca_parse_encoding_name (const <link linkend="char">char</link> *name);</programlisting> -<para> -Transofrms encoding specification charset/surface into numeric <link linkend="EncaEncoding"><type>EncaEncoding</type></link>. -</para> -<para> -When the charset name is not recognized, surfaces are not parsed at all and -<link linkend="ENCA-CS-UNKNOWN--CAPS"><type>ENCA_CS_UNKNOWN</type></link> is returned as charset. However, unrecognized surfaces are -considered only a minor problem causing <link linkend="ENCA-SURFACE-UNKNOWN--CAPS"><literal>ENCA_SURFACE_UNKNOWN</literal></link> flag to be -set in the result, beside recognized surface flags.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>name</parameter> :</term> -<listitem><simpara> An encoding specification. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The charset/surface pair. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-natural-surface" role="function"> -<title>enca_charset_natural_surface ()</title> -<indexterm zone="enca-charset-natural-surface"><primary sortas="enca_charset_natural_surface">enca_charset_natural_surface</primary></indexterm><programlisting><link linkend="EncaSurface">EncaSurface</link> enca_charset_natural_surface (<link linkend="int">int</link> charset);</programlisting> -<para> -Returns natural surface of a charset.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset</parameter> :</term> -<listitem><simpara> A charset. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The requested charset natural surface (called `implied' in recode), - zero for invalid <parameter>charset</parameter> or for charsets with no natural surface. - - Natrual surface is the surface one expects for a given charset -- - e.g. CRLF EOLs for IBM/Microsoft charsets, CR EOLs for Macintosh - charsets and LF EOLs for ISO/Unix charsets. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-properties" role="function"> -<title>enca_charset_properties ()</title> -<indexterm zone="enca-charset-properties"><primary sortas="enca_charset_properties">enca_charset_properties</primary></indexterm><programlisting><link linkend="EncaCharsetFlags">EncaCharsetFlags</link> enca_charset_properties (<link linkend="int">int</link> charset);</programlisting> -<para> -Returns charset properties.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset</parameter> :</term> -<listitem><simpara> A charset. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The requested charset properties; zero for invalid <parameter>charset</parameter>. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-7bit" role="macro"> -<title>enca_charset_is_7bit()</title> -<indexterm zone="enca-charset-is-7bit"><primary sortas="enca_charset_is_7bit">enca_charset_is_7bit</primary></indexterm><programlisting>#define enca_charset_is_7bit(cs)</programlisting> -<para> -Expands to nonzero when characters are represented with 7bit characters.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-8bit" role="macro"> -<title>enca_charset_is_8bit()</title> -<indexterm zone="enca-charset-is-8bit"><primary sortas="enca_charset_is_8bit">enca_charset_is_8bit</primary></indexterm><programlisting>#define enca_charset_is_8bit(cs)</programlisting> -<para> -Expands to nonzero when characters are represented with bytes.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-16bit" role="macro"> -<title>enca_charset_is_16bit()</title> -<indexterm zone="enca-charset-is-16bit"><primary sortas="enca_charset_is_16bit">enca_charset_is_16bit</primary></indexterm><programlisting>#define enca_charset_is_16bit(cs)</programlisting> -<para> -Expands to nonzero when characters are represented with 2byte words.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-32bit" role="macro"> -<title>enca_charset_is_32bit()</title> -<indexterm zone="enca-charset-is-32bit"><primary sortas="enca_charset_is_32bit">enca_charset_is_32bit</primary></indexterm><programlisting>#define enca_charset_is_32bit(cs)</programlisting> -<para> -Expands to nonzero when characters are represented with 4byte words.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-fixed" role="macro"> -<title>enca_charset_is_fixed()</title> -<indexterm zone="enca-charset-is-fixed"><primary sortas="enca_charset_is_fixed">enca_charset_is_fixed</primary></indexterm><programlisting>#define enca_charset_is_fixed(cs)</programlisting> -<para> -Expands to nonzero when one characters consists of one fundamental piece.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-variable" role="macro"> -<title>enca_charset_is_variable()</title> -<indexterm zone="enca-charset-is-variable"><primary sortas="enca_charset_is_variable">enca_charset_is_variable</primary></indexterm><programlisting>#define enca_charset_is_variable(cs)</programlisting> -<para> -Expands to nonzero when one character consists of variable number of -fundamental pieces.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-binary" role="macro"> -<title>enca_charset_is_binary()</title> -<indexterm zone="enca-charset-is-binary"><primary sortas="enca_charset_is_binary">enca_charset_is_binary</primary></indexterm><programlisting>#define enca_charset_is_binary(cs)</programlisting> -<para> -Expands to nonzero when charset is binary from ASCII viewpoint.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-regular" role="macro"> -<title>enca_charset_is_regular()</title> -<indexterm zone="enca-charset-is-regular"><primary sortas="enca_charset_is_regular">enca_charset_is_regular</primary></indexterm><programlisting>#define enca_charset_is_regular(cs)</programlisting> -<para> -Expands to nonzero when charset is language dependent (8bit) charset.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-multibyte" role="macro"> -<title>enca_charset_is_multibyte()</title> -<indexterm zone="enca-charset-is-multibyte"><primary sortas="enca_charset_is_multibyte">enca_charset_is_multibyte</primary></indexterm><programlisting>#define enca_charset_is_multibyte(cs)</programlisting> -<para> -Expands to nonzero when charset is multibyte.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charset-is-known" role="macro"> -<title>enca_charset_is_known()</title> -<indexterm zone="enca-charset-is-known"><primary sortas="enca_charset_is_known">enca_charset_is_known</primary></indexterm><programlisting>#define enca_charset_is_known(cs)</programlisting> -<para> -Expands to nonzero when the charset is known (i.e. it's not -ENCA_CS_UNKNOWN).</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> Charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> - -</refsect1> - - - - -</refentry> diff --git a/lib/enca/devel-docs/xml/enums.xml b/lib/enca/devel-docs/xml/enums.xml deleted file mode 100644 index ca4e66aedc..0000000000 --- a/lib/enca/devel-docs/xml/enums.xml +++ /dev/null @@ -1,387 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<refentry id="libenca-Typedefs-and-Constants"> -<refmeta> -<refentrytitle role="top_of_page" id="libenca-Typedefs-and-Constants.top_of_page">Typedefs and Constants</refentrytitle> -<manvolnum>3</manvolnum> -<refmiscinfo>LIBENCA Library</refmiscinfo> -</refmeta> - -<refnamediv> -<refname>Typedefs and Constants</refname> -<refpurpose>Enca library typedefs, enums and constants.</refpurpose> -</refnamediv> - -<refsynopsisdiv id="libenca-Typedefs-and-Constants.synopsis" role="synopsis"> -<title role="synopsis.title">Synopsis</title> - -<synopsis> - <link linkend="EncaEncoding">EncaEncoding</link>; -#define <link linkend="ENCA-CS-UNKNOWN--CAPS">ENCA_CS_UNKNOWN</link> -enum <link linkend="EncaSurface">EncaSurface</link>; -enum <link linkend="EncaCharsetFlags">EncaCharsetFlags</link>; -enum <link linkend="EncaNameStyle">EncaNameStyle</link>; -enum <link linkend="EncaErrno">EncaErrno</link>; -#define <link linkend="ENCA-NOT-A-CHAR--CAPS">ENCA_NOT_A_CHAR</link> -</synopsis> -</refsynopsisdiv> - - - - - - - - - -<refsect1 id="libenca-Typedefs-and-Constants.description" role="desc"> -<title role="desc.title">Description</title> -<para> -</para> -</refsect1> - -<refsect1 id="libenca-Typedefs-and-Constants.details" role="details"> -<title role="details.title">Details</title> -<refsect2 id="EncaEncoding" role="struct"> -<title>EncaEncoding</title> -<indexterm zone="EncaEncoding"><primary sortas="EncaEncoding">EncaEncoding</primary></indexterm><programlisting>typedef struct { - int charset; EncaSurface surface; -} EncaEncoding; -</programlisting> -<para> -Encoding, i.e. charset and surface. -</para> -<para> -This is what <link linkend="enca-analyse"><function>enca_analyse()</function></link> and <link linkend="enca-analyse-const"><function>enca_analyse_const()</function></link> return. -</para> -<para> -The <parameter>charset</parameter> field is an opaque numerical charset identifier, which has no -meaning outside Enca library. -You will probably want to use it only as <link linkend="enca-charset-name"><function>enca_charset_name()</function></link> argument. -It is only guaranteed not to change meaning -during program execution time; change of its interpretation (e.g. due to -addition of new charsets) is not considered API change. -</para> -<para> -The <parameter>surface</parameter> field is a combination of <link linkend="EncaSurface"><type>EncaSurface</type></link> flags. You may want -to ignore it completely; you should use <link linkend="enca-set-interpreted-surfaces"><function>enca_set_interpreted_surfaces()</function></link> -to disable weird surfaces then.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term><link linkend="int">int</link> <structfield>charset</structfield>;</term> -<listitem><simpara> Numeric charset identifier. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaSurface">EncaSurface</link> <structfield>surface</structfield>;</term> -<listitem><simpara> Surface flags. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="ENCA-CS-UNKNOWN--CAPS" role="macro"> -<title>ENCA_CS_UNKNOWN</title> -<indexterm zone="ENCA-CS-UNKNOWN--CAPS"><primary sortas="ENCA_CS_UNKNOWN">ENCA_CS_UNKNOWN</primary></indexterm><programlisting>#define ENCA_CS_UNKNOWN (-1) -</programlisting> -<para> -Unknown character set id. -</para> -<para> -Use <link linkend="enca-charset-is-known"><function>enca_charset_is_known()</function></link> to check for unknown charset instead of direct -comparsion.</para> -<para> -</para></refsect2> -<refsect2 id="EncaSurface" role="enum"> -<title>enum EncaSurface</title> -<indexterm zone="EncaSurface"><primary sortas="EncaSurface">EncaSurface</primary></indexterm><programlisting>typedef enum { /*< flags >*/ - ENCA_SURFACE_EOL_CR = 1 << 0, - ENCA_SURFACE_EOL_LF = 1 << 1, - ENCA_SURFACE_EOL_CRLF = 1 << 2, - ENCA_SURFACE_EOL_MIX = 1 << 3, - ENCA_SURFACE_EOL_BIN = 1 << 4, - ENCA_SURFACE_MASK_EOL = (ENCA_SURFACE_EOL_CR - | ENCA_SURFACE_EOL_LF - | ENCA_SURFACE_EOL_CRLF - | ENCA_SURFACE_EOL_MIX - | ENCA_SURFACE_EOL_BIN), - ENCA_SURFACE_PERM_21 = 1 << 5, - ENCA_SURFACE_PERM_4321 = 1 << 6, - ENCA_SURFACE_PERM_MIX = 1 << 7, - ENCA_SURFACE_MASK_PERM = (ENCA_SURFACE_PERM_21 - | ENCA_SURFACE_PERM_4321 - | ENCA_SURFACE_PERM_MIX), - ENCA_SURFACE_QP = 1 << 8, - ENCA_SURFACE_REMOVE = 1 << 13, - ENCA_SURFACE_UNKNOWN = 1 << 14, - ENCA_SURFACE_MASK_ALL = (ENCA_SURFACE_MASK_EOL - | ENCA_SURFACE_MASK_PERM - | ENCA_SURFACE_QP - | ENCA_SURFACE_REMOVE) -} EncaSurface; -</programlisting> -<para> -Surface flags.</para> -<para> -</para><variablelist role="enum"> -<varlistentry id="ENCA-SURFACE-EOL-CR--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_EOL_CR</literal></term> -<listitem><simpara> End-of-lines are represented with CR's. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-EOL-LF--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_EOL_LF</literal></term> -<listitem><simpara> End-of-lines are represented with LF's. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-EOL-CRLF--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_EOL_CRLF</literal></term> -<listitem><simpara> End-of-lines are represented with CRLF's. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-EOL-MIX--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_EOL_MIX</literal></term> -<listitem><simpara> Several end-of-line types, mixed. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-EOL-BIN--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_EOL_BIN</literal></term> -<listitem><simpara> End-of-line concept not applicable (binary data). -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-MASK-EOL--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_MASK_EOL</literal></term> -<listitem><simpara> Mask for end-of-line surfaces. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-PERM-21--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_PERM_21</literal></term> -<listitem><simpara> Odd and even bytes swapped. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-PERM-4321--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_PERM_4321</literal></term> -<listitem><simpara> Reversed byte sequence in 4byte words. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-PERM-MIX--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_PERM_MIX</literal></term> -<listitem><simpara> Chunks with both endianess, concatenated. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-MASK-PERM--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_MASK_PERM</literal></term> -<listitem><simpara> Mask for permutation surfaces. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-QP--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_QP</literal></term> -<listitem><simpara> Quoted printables. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-REMOVE--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_REMOVE</literal></term> -<listitem><simpara> Recode `remove' surface. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-UNKNOWN--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_UNKNOWN</literal></term> -<listitem><simpara> Unknown surface. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-SURFACE-MASK-ALL--CAPS" role="constant"> -<term><literal>ENCA_SURFACE_MASK_ALL</literal></term> -<listitem><simpara> Mask for all bits, withnout <link linkend="ENCA-SURFACE-UNKNOWN--CAPS"><type>ENCA_SURFACE_UNKNOWN</type></link>. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaCharsetFlags" role="enum"> -<title>enum EncaCharsetFlags</title> -<indexterm zone="EncaCharsetFlags"><primary sortas="EncaCharsetFlags">EncaCharsetFlags</primary></indexterm><programlisting>typedef enum { /*< flags >*/ - ENCA_CHARSET_7BIT = 1 << 0, - ENCA_CHARSET_8BIT = 1 << 1, - ENCA_CHARSET_16BIT = 1 << 2, - ENCA_CHARSET_32BIT = 1 << 3, - ENCA_CHARSET_FIXED = 1 << 4, - ENCA_CHARSET_VARIABLE = 1 << 5, - ENCA_CHARSET_BINARY = 1 << 6, - ENCA_CHARSET_REGULAR = 1 << 7, - ENCA_CHARSET_MULTIBYTE = 1 << 8 -} EncaCharsetFlags; -</programlisting> -<para> -Charset properties. -</para> -<para> -Flags <link linkend="ENCA-CHARSET-7BIT--CAPS"><literal>ENCA_CHARSET_7BIT</literal></link>, <link linkend="ENCA-CHARSET-8BIT--CAPS"><literal>ENCA_CHARSET_8BIT</literal></link>, <link linkend="ENCA-CHARSET-16BIT--CAPS"><literal>ENCA_CHARSET_16BIT</literal></link>, -<link linkend="ENCA-CHARSET-32BIT--CAPS"><literal>ENCA_CHARSET_32BIT</literal></link> tell how many bits a `fundamental piece' consists of. -This is different from bits per character; r.g. UTF-8 consists of 8bit -pieces (bytes), but character can be composed from 1 to 6 of them.</para> -<para> -</para><variablelist role="enum"> -<varlistentry id="ENCA-CHARSET-7BIT--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_7BIT</literal></term> -<listitem><simpara> Characters are represented with 7bit characters. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-8BIT--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_8BIT</literal></term> -<listitem><simpara> Characters are represented with bytes. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-16BIT--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_16BIT</literal></term> -<listitem><simpara> Characters are represented with 2byte words. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-32BIT--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_32BIT</literal></term> -<listitem><simpara> Characters are represented with 4byte words. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-FIXED--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_FIXED</literal></term> -<listitem><simpara> One characters consists of one fundamental piece. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-VARIABLE--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_VARIABLE</literal></term> -<listitem><simpara> One character consists of variable number of -fundamental pieces. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-BINARY--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_BINARY</literal></term> -<listitem><simpara> Charset is binary from ASCII viewpoint. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-REGULAR--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_REGULAR</literal></term> -<listitem><simpara> Language dependent (8bit) charset. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-CHARSET-MULTIBYTE--CAPS" role="constant"> -<term><literal>ENCA_CHARSET_MULTIBYTE</literal></term> -<listitem><simpara> Multibyte charset. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaNameStyle" role="enum"> -<title>enum EncaNameStyle</title> -<indexterm zone="EncaNameStyle"><primary sortas="EncaNameStyle">EncaNameStyle</primary></indexterm><programlisting>typedef enum { - ENCA_NAME_STYLE_ENCA, - ENCA_NAME_STYLE_RFC1345, - ENCA_NAME_STYLE_CSTOCS, - ENCA_NAME_STYLE_ICONV, - ENCA_NAME_STYLE_HUMAN, - ENCA_NAME_STYLE_MIME -} EncaNameStyle; -</programlisting> -<para> -Charset naming styles and conventions.</para> -<para> -</para><variablelist role="enum"> -<varlistentry id="ENCA-NAME-STYLE-ENCA--CAPS" role="constant"> -<term><literal>ENCA_NAME_STYLE_ENCA</literal></term> -<listitem><simpara> Default, implicit charset name in Enca. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-NAME-STYLE-RFC1345--CAPS" role="constant"> -<term><literal>ENCA_NAME_STYLE_RFC1345</literal></term> -<listitem><simpara> RFC 1345 or otherwise canonical charset name. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-NAME-STYLE-CSTOCS--CAPS" role="constant"> -<term><literal>ENCA_NAME_STYLE_CSTOCS</literal></term> -<listitem><simpara> Cstocs charset name (may not exist). -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-NAME-STYLE-ICONV--CAPS" role="constant"> -<term><literal>ENCA_NAME_STYLE_ICONV</literal></term> -<listitem><simpara> Iconv charset name (may not exist). -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-NAME-STYLE-HUMAN--CAPS" role="constant"> -<term><literal>ENCA_NAME_STYLE_HUMAN</literal></term> -<listitem><simpara> Human comprehensible description. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-NAME-STYLE-MIME--CAPS" role="constant"> -<term><literal>ENCA_NAME_STYLE_MIME</literal></term> -<listitem><simpara> Preferred MIME name (may not exist). -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaErrno" role="enum"> -<title>enum EncaErrno</title> -<indexterm zone="EncaErrno"><primary sortas="EncaErrno">EncaErrno</primary></indexterm><programlisting>typedef enum { - ENCA_EOK = 0, - ENCA_EINVALUE, - ENCA_EEMPTY, - ENCA_EFILTERED, - ENCA_ENOCS8, - ENCA_ESIGNIF, - ENCA_EWINNER, - ENCA_EGARBAGE -} EncaErrno; -</programlisting> -<para> -Error codes.</para> -<para> -</para><variablelist role="enum"> -<varlistentry id="ENCA-EOK--CAPS" role="constant"> -<term><literal>ENCA_EOK</literal></term> -<listitem><simpara> OK. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-EINVALUE--CAPS" role="constant"> -<term><literal>ENCA_EINVALUE</literal></term> -<listitem><simpara> Invalid value (usually of an option). -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-EEMPTY--CAPS" role="constant"> -<term><literal>ENCA_EEMPTY</literal></term> -<listitem><simpara> Sample is empty. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-EFILTERED--CAPS" role="constant"> -<term><literal>ENCA_EFILTERED</literal></term> -<listitem><simpara> After filtering, (almost) nothing remained. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-ENOCS8--CAPS" role="constant"> -<term><literal>ENCA_ENOCS8</literal></term> -<listitem><simpara> Mulitibyte tests failed and language contains no 8bit charsets. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-ESIGNIF--CAPS" role="constant"> -<term><literal>ENCA_ESIGNIF</literal></term> -<listitem><simpara> Too few significant characters. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-EWINNER--CAPS" role="constant"> -<term><literal>ENCA_EWINNER</literal></term> -<listitem><simpara> No clear winner. -</simpara></listitem> -</varlistentry> -<varlistentry id="ENCA-EGARBAGE--CAPS" role="constant"> -<term><literal>ENCA_EGARBAGE</literal></term> -<listitem><simpara> Sample is garbage. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="ENCA-NOT-A-CHAR--CAPS" role="macro"> -<title>ENCA_NOT_A_CHAR</title> -<indexterm zone="ENCA-NOT-A-CHAR--CAPS"><primary sortas="ENCA_NOT_A_CHAR">ENCA_NOT_A_CHAR</primary></indexterm><programlisting>#define ENCA_NOT_A_CHAR 0xffff -</programlisting> -<para> -Not-a-character in unicode tables.</para> -<para> -</para></refsect2> - -</refsect1> - - - - -</refentry> diff --git a/lib/enca/devel-docs/xml/internal.xml b/lib/enca/devel-docs/xml/internal.xml deleted file mode 100644 index 37fafc197a..0000000000 --- a/lib/enca/devel-docs/xml/internal.xml +++ /dev/null @@ -1,1328 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE refentry PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<refentry id="libenca-Internal-Functions"> -<refmeta> -<refentrytitle role="top_of_page" id="libenca-Internal-Functions.top_of_page">internal</refentrytitle> -<manvolnum>3</manvolnum> -<refmiscinfo>LIBENCA Library</refmiscinfo> -</refmeta> - -<refnamediv> -<refname>internal</refname> -<refpurpose><para> -internal functions -</para></refpurpose> -</refnamediv> - -<refsynopsisdiv id="libenca-Internal-Functions.synopsis" role="synopsis"> -<title role="synopsis.title">Synopsis</title> - -<synopsis> - <link linkend="EncaAnalyserOptions">EncaAnalyserOptions</link>; - <link linkend="EncaAnalyserState">EncaAnalyserState</link>; - <link linkend="EncaCharsetInfo">EncaCharsetInfo</link>; - <link linkend="EncaLanguageInfo">EncaLanguageInfo</link>; - <link linkend="EncaLanguageHookData1CS">EncaLanguageHookData1CS</link>; - <link linkend="EncaLanguageHookDataEOL">EncaLanguageHookDataEOL</link>; -<link linkend="int">int</link> (<link linkend="EncaHookFunc">*EncaHookFunc</link>) (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="int">int</link> (<link linkend="EncaGuessFunc">*EncaGuessFunc</link>) (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); - <link linkend="EncaUTFCheckData">EncaUTFCheckData</link>; -#define <link linkend="ELEMENTS--CAPS">ELEMENTS</link> (array) -#define <link linkend="MAKE-HOOK-LINE--CAPS">MAKE_HOOK_LINE</link> (name) -#define <link linkend="EPSILON--CAPS">EPSILON</link> -#define <link linkend="FILL-NONLETTER--CAPS">FILL_NONLETTER</link> -#define <link linkend="LF--CAPS">LF</link> -#define <link linkend="CR--CAPS">CR</link> -<link linkend="void">void</link>* <link linkend="enca-malloc">enca_malloc</link> (<link linkend="size-t">size_t</link> size); -<link linkend="void">void</link>* <link linkend="enca-realloc">enca_realloc</link> (<link linkend="void">void</link> *ptr, - <link linkend="size-t">size_t</link> size); -#define <link linkend="enca-free">enca_free</link> (ptr) -#define <link linkend="NEW--CAPS">NEW</link> (type,n) -#define <link linkend="RENEW--CAPS">RENEW</link> (ptr,type,n) -<link linkend="char">char</link>* <link linkend="enca-strdup">enca_strdup</link> (const <link linkend="char">char</link> *s); -const <link linkend="char">char</link>* <link linkend="enca-strstr">enca_strstr</link> (const <link linkend="char">char</link> *haystack, - const <link linkend="char">char</link> *needle); -<link linkend="char">char</link>* <link linkend="enca-stpcpy">enca_stpcpy</link> (<link linkend="char">char</link> *dest, - const <link linkend="char">char</link> *src); -<link linkend="char">char</link>* <link linkend="enca-strconcat">enca_strconcat</link> (const <link linkend="char">char</link> *str, - ...); -<link linkend="char">char</link>* <link linkend="enca-strappend">enca_strappend</link> (<link linkend="char">char</link> *str, - ...); -#define <link linkend="enca-csname">enca_csname</link> (cs) -<link linkend="int">int</link> <link linkend="enca-name-to-charset">enca_name_to_charset</link> (const <link linkend="char">char</link> *csname); -<link linkend="EncaSurface">EncaSurface</link> <link linkend="enca-name-to-surface">enca_name_to_surface</link> (const <link linkend="char">char</link> *sname); -<link linkend="int">int</link> <link linkend="enca-language-init">enca_language_init</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - const <link linkend="char">char</link> *langname); -<link linkend="void">void</link> <link linkend="enca-language-destroy">enca_language_destroy</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="double">double</link>* <link linkend="enca-get-charset-similarity-matrix">enca_get_charset_similarity_matrix</link> (const <link linkend="EncaLanguageInfo">EncaLanguageInfo</link> *lang); -<link linkend="int">int</link> <link linkend="enca-charsets-subset-identical">enca_charsets_subset_identical</link> (<link linkend="int">int</link> charset1, - <link linkend="int">int</link> charset2, - const <link linkend="size-t">size_t</link> *counts); -<link linkend="size-t">size_t</link> <link linkend="enca-filter-boxdraw">enca_filter_boxdraw</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - unsigned <link linkend="char">char</link> fill_char); -<link linkend="int">int</link> <link linkend="enca-language-hook-ncs">enca_language_hook_ncs</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - <link linkend="size-t">size_t</link> ncs, - <link linkend="EncaLanguageHookData1CS">EncaLanguageHookData1CS</link> *hookdata); -<link linkend="int">int</link> <link linkend="enca-language-hook-eol">enca_language_hook_eol</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - <link linkend="size-t">size_t</link> ncs, - <link linkend="EncaLanguageHookDataEOL">EncaLanguageHookDataEOL</link> *hookdata); -<link linkend="void">void</link> <link linkend="enca-guess-init">enca_guess_init</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="void">void</link> <link linkend="enca-guess-destroy">enca_guess_destroy</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="EncaSurface">EncaSurface</link> <link linkend="enca-eol-surface">enca_eol_surface</link> (unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size, - const <link linkend="size-t">size_t</link> *counts); -<link linkend="void">void</link> <link linkend="enca-find-max-sec">enca_find_max_sec</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-BE--CAPS">ENCA_LANGUAGE_BE</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-BG--CAPS">ENCA_LANGUAGE_BG</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-CS--CAPS">ENCA_LANGUAGE_CS</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-ET--CAPS">ENCA_LANGUAGE_ET</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-HR--CAPS">ENCA_LANGUAGE_HR</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-HU--CAPS">ENCA_LANGUAGE_HU</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-LT--CAPS">ENCA_LANGUAGE_LT</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-LV--CAPS">ENCA_LANGUAGE_LV</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-PL--CAPS">ENCA_LANGUAGE_PL</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-RU--CAPS">ENCA_LANGUAGE_RU</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-SK--CAPS">ENCA_LANGUAGE_SK</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-SL--CAPS">ENCA_LANGUAGE_SL</link>; -extern const EncaLanguageInfo <link linkend="ENCA-LANGUAGE-UK--CAPS">ENCA_LANGUAGE_UK</link>; -<link linkend="void">void</link> <link linkend="enca-double-utf8-init">enca_double_utf8_init</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="void">void</link> <link linkend="enca-double-utf8-destroy">enca_double_utf8_destroy</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="void">void</link> <link linkend="enca-pair-init">enca_pair_init</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="void">void</link> <link linkend="enca-pair-destroy">enca_pair_destroy</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -<link linkend="int">int</link> <link linkend="enca-pair-analyse">enca_pair_analyse</link> (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser); -</synopsis> -</refsynopsisdiv> - - - - - - - - - -<refsect1 id="libenca-Internal-Functions.description" role="desc"> -<title role="desc.title">Description</title> -<para> -Do not use outside Enca library. -</para> -</refsect1> - -<refsect1 id="libenca-Internal-Functions.details" role="details"> -<title role="details.title">Details</title> -<refsect2 id="EncaAnalyserOptions" role="struct"> -<title>EncaAnalyserOptions</title> -<indexterm zone="EncaAnalyserOptions"><primary sortas="EncaAnalyserOptions">EncaAnalyserOptions</primary></indexterm><programlisting>typedef struct { - int const_buffer; - size_t min_chars; - double threshold; - int multibyte_enabled; - int interpreted_surfaces; - int ambiguous_mode; - int filtering; - int test_garbageness; - int termination_strictness; -} EncaAnalyserOptions; -</programlisting> -<para> -Analyser options, a part of analyser state.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term><link linkend="int">int</link> <structfield>const_buffer</structfield>;</term> -<listitem><simpara> Treat buffer as const? Otherwise its content can be, - and probably will be, modified. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>min_chars</structfield>;</term> -<listitem><simpara> Minimal number significant characters. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="double">double</link> <structfield>threshold</structfield>;</term> -<listitem><simpara> Minimal ratio between winner and the second. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>multibyte_enabled</structfield>;</term> -<listitem><simpara> Check for multibyte encodings? -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>interpreted_surfaces</structfield>;</term> -<listitem><simpara> Allow surfaces causing fundamental reinterpretation? -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>ambiguous_mode</structfield>;</term> -<listitem><simpara> Ambiguous mode? -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>filtering</structfield>;</term> -<listitem><simpara> Allow binary and box-drawing filters? -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>test_garbageness</structfield>;</term> -<listitem><simpara> Do test garbageness? -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>termination_strictness</structfield>;</term> -<listitem><simpara> Disallow broken multibyte sequences at buffer end? -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaAnalyserState" role="struct"> -<title>EncaAnalyserState</title> -<indexterm zone="EncaAnalyserState"><primary sortas="EncaAnalyserState">EncaAnalyserState</primary></indexterm><programlisting>typedef struct { - /* Language data. */ - const EncaLanguageInfo *lang; - size_t ncharsets; - int *charsets; - /* Analyser state. */ - EncaErrno gerrno; - size_t size; - unsigned char *buffer; - EncaEncoding result; - size_t *counts; - size_t bin; - size_t up; - double *ratings; - size_t *order; - size_t size2; - unsigned char *buffer2; - /* Double-UTF-8 data. */ - EncaUTFCheckData *utfch; - int *utfbuf; - /* Pair frequency data */ - unsigned char *pair2bits; - size_t *bitcounts; - size_t *pairratings; - /* LCUC data XXX: unused (yet) */ - size_t *lcbits; - size_t *ucbits; - /* Options. */ - EncaAnalyserOptions options; -} EncaAnalyserState; -</programlisting> -<para> -The internal analyser state. -</para> -<para> -Passed as an opaque object (`this') to analyser calls.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term>const <link linkend="EncaLanguageInfo">EncaLanguageInfo</link> *<structfield>lang</structfield>;</term> -<listitem><simpara> Language informations. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>ncharsets</structfield>;</term> -<listitem><simpara> Number of 8bit charsets in this language. - (Equal to <parameter>lang</parameter>->ncharsets.) -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> *<structfield>charsets</structfield>;</term> -<listitem><simpara> 8bit charset id's [<parameter>ncharsets</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaErrno">EncaErrno</link> <structfield>gerrno</structfield>;</term> -<listitem><simpara> Guessing gerrno. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>size</structfield>;</term> -<listitem><simpara> Size of buffer. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>unsigned <link linkend="char">char</link> *<structfield>buffer</structfield>;</term> -<listitem><simpara> Buffer whose encoding is to be detected [<parameter>size</parameter>]. - (Owned by outer world.) -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaEncoding">EncaEncoding</link> <structfield>result</structfield>;</term> -<listitem><simpara> Result returned to caller. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> *<structfield>counts</structfield>;</term> -<listitem><simpara> Character counts [0x100]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>bin</structfield>;</term> -<listitem><simpara> Number of `binary' characters. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>up</structfield>;</term> -<listitem><simpara> Number of 8bit characters. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="double">double</link> *<structfield>ratings</structfield>;</term> -<listitem><simpara> 8bit charset ratings [<parameter>ncharsets</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> *<structfield>order</structfield>;</term> -<listitem><simpara> Charset indices (not id's) sorted by ratings in descending order - [ncharsets]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>size2</structfield>;</term> -<listitem><simpara> Size of buffer2. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>unsigned <link linkend="char">char</link> *<structfield>buffer2</structfield>;</term> -<listitem><simpara> A temporary secondary buffer [<parameter>size2</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaUTFCheckData">EncaUTFCheckData</link> *<structfield>utfch</structfield>;</term> -<listitem><simpara> Double-UTF-8 test data [<parameter>ncharsets</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> *<structfield>utfbuf</structfield>;</term> -<listitem><simpara> Double-UTF-8 buffer for various UCS-2 character counting [0x10000]. - (Magic: see <link linkend="mark-scratch-buffer"><function>mark_scratch_buffer()</function></link> for description.) -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>unsigned <link linkend="char">char</link> *<structfield>pair2bits</structfield>;</term> -<listitem><simpara> Character pair map to charsets [0x100000] (indexed - 0x100*first + second). Each bit corresponds to one charset, - when set, the pair is `good' for the given charset. The - type is char, so it breaks for <parameter>ncharsets</parameter> > 8, but it should - not be accessed from outer world, so it can be easily enlarged - to more bits. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> *<structfield>bitcounts</structfield>;</term> -<listitem><simpara> Counts for each possible bit combinations in <parameter>pair2bits</parameter> - [0x1 << ncharsets]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> *<structfield>pairratings</structfield>;</term> -<listitem><simpara> Counts of `good' pairs per charset [<parameter>ncharsets</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> *<structfield>lcbits</structfield>;</term> -<listitem><simpara> If a character is lowercase in some charset, correspinding bit - is set [0x100]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> *<structfield>ucbits</structfield>;</term> -<listitem><simpara> If a character is uppercase in some charset, correspinding bit - is set [0x100]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaAnalyserOptions">EncaAnalyserOptions</link> <structfield>options</structfield>;</term> -<listitem><simpara> Analyser options. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaCharsetInfo" role="struct"> -<title>EncaCharsetInfo</title> -<indexterm zone="EncaCharsetInfo"><primary sortas="EncaCharsetInfo">EncaCharsetInfo</primary></indexterm><programlisting>typedef struct { - int enca; - int rfc1345; - int cstocs; - int iconv; - int mime; - const char *human; - unsigned int flags; - unsigned int nsurface; -} EncaCharsetInfo; -</programlisting> -<para> -General charset informnations. -</para> -<para> -All the <link linkend="int"><type>int</type></link> fields are indices in <link linkend="ALIAS-LIST--CAPS"><type>ALIAS_LIST</type></link>[].</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term><link linkend="int">int</link> <structfield>enca</structfield>;</term> -<listitem><simpara> Default, implicit name in enca. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>rfc1345</structfield>;</term> -<listitem><simpara> RFC1345 charset name. - (For charsets not in RFC1345, some canonical name is invented.) -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>cstocs</structfield>;</term> -<listitem><simpara> Cstocs charset name or -1. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>iconv</structfield>;</term> -<listitem><simpara> Iconv charset name or -1. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>mime</structfield>;</term> -<listitem><simpara> Preferred MIME charset name or -1. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>const <link linkend="char">char</link> *<structfield>human</structfield>;</term> -<listitem><simpara> Human comprehensible description. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>unsigned <link linkend="int">int</link> <structfield>flags</structfield>;</term> -<listitem><simpara> Charset properties (7bit, 8bit, multibyte, ...). -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>unsigned <link linkend="int">int</link> <structfield>nsurface</structfield>;</term> -<listitem><simpara> Natural surface (`implied' in recode). -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaLanguageInfo" role="struct"> -<title>EncaLanguageInfo</title> -<indexterm zone="EncaLanguageInfo"><primary sortas="EncaLanguageInfo">EncaLanguageInfo</primary></indexterm><programlisting>typedef struct { - const char *name; - const char *humanname; - size_t ncharsets; - const char *const *csnames; - const unsigned short int *const *weights; - const unsigned short int *significant; - const unsigned char *const *letters; - const unsigned char **const *pairs; - long int weight_sum; - EncaHookFunc hook; - EncaHookFunc eolhook; - EncaHookFunc lcuchook; - EncaHookFunc ratinghook; -} EncaLanguageInfo; -</programlisting> -<para> -Language specific data.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term>const <link linkend="char">char</link> *<structfield>name</structfield>;</term> -<listitem><simpara> Language name, or more precisely, locale name. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>const <link linkend="char">char</link> *<structfield>humanname</structfield>;</term> -<listitem><simpara> Normal human-readable [English] language name. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>ncharsets</structfield>;</term> -<listitem><simpara> Number of charsets in this language. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>const <link linkend="char">char</link> *const  *<structfield>csnames</structfield>;</term> -<listitem><simpara> Charset names [<parameter>ncharsets</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term>long <link linkend="int">int</link> <structfield>weight_sum</structfield>;</term> -<listitem><simpara> Sum of all <parameter>weights</parameter> (is the same for all charsets). -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaHookFunc">EncaHookFunc</link> <structfield>hook</structfield>;</term> -<listitem><simpara> Hook function (deciding hard cases). -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaHookFunc">EncaHookFunc</link> <structfield>eolhook</structfield>;</term> -<listitem><simpara> EOL hook function (deciding ambiguous cases based on EOL type). -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaHookFunc">EncaHookFunc</link> <structfield>lcuchook</structfield>;</term> -<listitem><simpara> -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaHookFunc">EncaHookFunc</link> <structfield>ratinghook</structfield>;</term> -<listitem><simpara> Helper to calculate ratings for weightingless languages. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaLanguageHookData1CS" role="struct"> -<title>EncaLanguageHookData1CS</title> -<indexterm zone="EncaLanguageHookData1CS"><primary sortas="EncaLanguageHookData1CS">EncaLanguageHookData1CS</primary></indexterm><programlisting>typedef struct { - const char *name; - size_t size; - const unsigned char *list; - size_t cs; -} EncaLanguageHookData1CS; -</programlisting> -<para> -Cointainer for data needed by <link linkend="enca-language-hook-ncs"><function>enca_language_hook_ncs()</function></link>.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term>const <link linkend="char">char</link> *<structfield>name</structfield>;</term> -<listitem><simpara> Charset name. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>size</structfield>;</term> -<listitem><simpara> Number of characters in <parameter>list</parameter>. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>cs</structfield>;</term> -<listitem><simpara> Charset number. This is an index in <parameter>analyser</parameter> arrays (like <parameter>charsets</parameter>), - NOT a charset id. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaLanguageHookDataEOL" role="struct"> -<title>EncaLanguageHookDataEOL</title> -<indexterm zone="EncaLanguageHookDataEOL"><primary sortas="EncaLanguageHookDataEOL">EncaLanguageHookDataEOL</primary></indexterm><programlisting>typedef struct { - const char *name; - EncaSurface eol; - size_t cs; -} EncaLanguageHookDataEOL; -</programlisting> -<para> -Cointainer for data needed by <link linkend="enca-language-hook-eol"><function>enca_language_hook_eol()</function></link>.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term>const <link linkend="char">char</link> *<structfield>name</structfield>;</term> -<listitem><simpara> Charset name. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="EncaSurface">EncaSurface</link> <structfield>eol</structfield>;</term> -<listitem><simpara> The corresponding <link linkend="EncaSurface"><type>EncaSurface</type></link> bit. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>cs</structfield>;</term> -<listitem><simpara> Charset number. This is an index in <parameter>analyser</parameter> arrays (like <parameter>charsets</parameter>), - NOT a charset id. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaHookFunc" role="function"> -<title>EncaHookFunc ()</title> -<indexterm zone="EncaHookFunc"><primary sortas="EncaHookFunc">EncaHookFunc</primary></indexterm><programlisting><link linkend="int">int</link> (*EncaHookFunc) (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Language hook function type. -</para> -<para> -Launches language specific hooks for a particular language.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser state whose charset ratings are to be modified. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero if charset ratigns have been actually modified, zero -otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaGuessFunc" role="function"> -<title>EncaGuessFunc ()</title> -<indexterm zone="EncaGuessFunc"><primary sortas="EncaGuessFunc">EncaGuessFunc</primary></indexterm><programlisting><link linkend="int">int</link> (*EncaGuessFunc) (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Special (multibyte) encoding check function type.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser state whose buffer should be checked. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero if analyser->result has been set, zero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="EncaUTFCheckData" role="struct"> -<title>EncaUTFCheckData</title> -<indexterm zone="EncaUTFCheckData"><primary sortas="EncaUTFCheckData">EncaUTFCheckData</primary></indexterm><programlisting>typedef struct { - double rating; - size_t size; - int result; - int *ucs2; - int *weights; -} EncaUTFCheckData; -</programlisting> -<para> -Data needed by double-UTF-8 check, per language charset.</para> -<para> -</para><variablelist role="struct"> -<varlistentry> -<term><link linkend="double">double</link> <structfield>rating</structfield>;</term> -<listitem><simpara> Total rating for this charset. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="size-t">size_t</link> <structfield>size</structfield>;</term> -<listitem><simpara> Number of UCS-2 characters. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> <structfield>result</structfield>;</term> -<listitem><simpara> Nonzero when the sample is probably Doubly-UTF-8 encoded from - this charset. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> *<structfield>ucs2</structfield>;</term> -<listitem><simpara> List of significant UCS-2 characters, in order [<parameter>size</parameter>]. -</simpara></listitem> -</varlistentry> -<varlistentry> -<term><link linkend="int">int</link> *<structfield>weights</structfield>;</term> -<listitem><simpara> Weights for double-UTF-8 check [<parameter>size</parameter>]. Positive means normal - UTF-8, negative doubly-encoded. -</simpara></listitem> -</varlistentry> -</variablelist></refsect2> -<refsect2 id="ELEMENTS--CAPS" role="macro"> -<title>ELEMENTS()</title> -<indexterm zone="ELEMENTS--CAPS"><primary sortas="ELEMENTS">ELEMENTS</primary></indexterm><programlisting>#define ELEMENTS(array) (sizeof(array)/sizeof((array)[0])) -</programlisting> -<para> -Compute the number of elements of a static array.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>array</parameter> :</term> -<listitem><simpara> An array whose size is to be computed. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> the number of elements. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="MAKE-HOOK-LINE--CAPS" role="macro"> -<title>MAKE_HOOK_LINE()</title> -<indexterm zone="MAKE-HOOK-LINE--CAPS"><primary sortas="MAKE_HOOK_LINE">MAKE_HOOK_LINE</primary></indexterm><programlisting>#define MAKE_HOOK_LINE(name)</programlisting> -<para> -Ugly code `beautifier' macro for language hooks.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>name</parameter> :</term> -<listitem><simpara> A charset name in C-style identifier suitable form. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="EPSILON--CAPS" role="macro"> -<title>EPSILON</title> -<indexterm zone="EPSILON--CAPS"><primary sortas="EPSILON">EPSILON</primary></indexterm><programlisting>#define EPSILON 0.000001 -</programlisting> -<para> -`Zero' for float comparsion (and to prevent division by zero, etc.).</para> -<para> -</para></refsect2> -<refsect2 id="FILL-NONLETTER--CAPS" role="macro"> -<title>FILL_NONLETTER</title> -<indexterm zone="FILL-NONLETTER--CAPS"><primary sortas="FILL_NONLETTER">FILL_NONLETTER</primary></indexterm><programlisting>#define FILL_NONLETTER '.' -</programlisting> -<para> -Replacement character for non-letters in pair frequencies.</para> -<para> -</para></refsect2> -<refsect2 id="LF--CAPS" role="macro"> -<title>LF</title> -<indexterm zone="LF--CAPS"><primary sortas="LF">LF</primary></indexterm><programlisting>#define LF ((unsigned char)'\n') -</programlisting> -<para> -Line feed character (End-of-line on Unix).</para> -<para> -</para></refsect2> -<refsect2 id="CR--CAPS" role="macro"> -<title>CR</title> -<indexterm zone="CR--CAPS"><primary sortas="CR">CR</primary></indexterm><programlisting>#define CR ((unsigned char)'\r') -</programlisting> -<para> -Carriage return character (End-of-line on Macintosh).</para> -<para> -</para></refsect2> -<refsect2 id="enca-malloc" role="function"> -<title>enca_malloc ()</title> -<indexterm zone="enca-malloc"><primary sortas="enca_malloc">enca_malloc</primary></indexterm><programlisting><link linkend="void">void</link>* enca_malloc (<link linkend="size-t">size_t</link> size);</programlisting> -<para> -Allocates memory, always successfully (when fails, aborts program).</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>size</parameter> :</term> -<listitem><simpara> The number of bytes to allocate. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Pointer to the newly allocated memory. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-realloc" role="function"> -<title>enca_realloc ()</title> -<indexterm zone="enca-realloc"><primary sortas="enca_realloc">enca_realloc</primary></indexterm><programlisting><link linkend="void">void</link>* enca_realloc (<link linkend="void">void</link> *ptr, - <link linkend="size-t">size_t</link> size);</programlisting> -<para> -Reallocates memory, always successfully (when fails, aborts program).</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>ptr</parameter> :</term> -<listitem><simpara> Pointer to block of previously allocated memory. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>size</parameter> :</term> -<listitem><simpara> The number of bytes to resize the block. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Pointer to the newly allocated memory, <link linkend="NULL--CAPS"><type>NULL</type></link> when <parameter>size</parameter> is zero. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-free" role="macro"> -<title>enca_free()</title> -<indexterm zone="enca-free"><primary sortas="enca_free">enca_free</primary></indexterm><programlisting>#define enca_free(ptr)</programlisting> -<para> -Frees memory pointed by <parameter>ptr</parameter> with <link linkend="free"><function>free()</function></link> hack and assigns it a safe value, -thus may be called more than once. -</para> -<para> -<parameter>ptr</parameter> MUST be l-value.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>ptr</parameter> :</term> -<listitem><simpara> Pointer to memory to free. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="NEW--CAPS" role="macro"> -<title>NEW()</title> -<indexterm zone="NEW--CAPS"><primary sortas="NEW">NEW</primary></indexterm><programlisting>#define NEW(type,n) ((type*)enca_malloc((n)*sizeof(type))) -</programlisting> -<para> -An <link linkend="enca-malloc"><function>enca_malloc()</function></link> wrapper.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>type</parameter> :</term> -<listitem><simpara> Data type to allocate. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>n</parameter> :</term> -<listitem><simpara> Number of elements to allocate. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Pointer to the newly allocated memory. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="RENEW--CAPS" role="macro"> -<title>RENEW()</title> -<indexterm zone="RENEW--CAPS"><primary sortas="RENEW">RENEW</primary></indexterm><programlisting>#define RENEW(ptr,type,n) ((type*)enca_realloc((ptr),(n)*sizeof(type))) -</programlisting> -<para> -An <link linkend="enca-realloc"><function>enca_realloc()</function></link> wrapper.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>ptr</parameter> :</term> -<listitem><simpara> Pointer to already allocate memory or <link linkend="NULL--CAPS"><type>NULL</type></link>. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>type</parameter> :</term> -<listitem><simpara> Data type to allocate. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>n</parameter> :</term> -<listitem><simpara> Number of elements to resize the memory to. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Pointer to the reallocated memory (or pointer safe to call <link linkend="free"><function>free()</function></link> -on when <parameter>n</parameter> is zero). -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-strdup" role="function"> -<title>enca_strdup ()</title> -<indexterm zone="enca-strdup"><primary sortas="enca_strdup">enca_strdup</primary></indexterm><programlisting><link linkend="char">char</link>* enca_strdup (const <link linkend="char">char</link> *s);</programlisting> -<para> -Duplicates string. -</para> -<para> -Will be defined as <link linkend="strdup"><function>strdup()</function></link> when system provides it.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>s</parameter> :</term> -<listitem><simpara> A string. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The newly allocated string copy. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-strstr" role="function"> -<title>enca_strstr ()</title> -<indexterm zone="enca-strstr"><primary sortas="enca_strstr">enca_strstr</primary></indexterm><programlisting>const <link linkend="char">char</link>* enca_strstr (const <link linkend="char">char</link> *haystack, - const <link linkend="char">char</link> *needle);</programlisting> -<para> -Finds occurence of a substring in a string. -</para> -<para> -Will be defined as <link linkend="strstr"><function>strstr()</function></link> when system provides it.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>haystack</parameter> :</term> -<listitem><simpara> A string where to search. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>needle</parameter> :</term> -<listitem><simpara> A string to find. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Pointer to the first occurence of <parameter>needle</parameter> in <parameter>haystack</parameter>; <link linkend="NULL--CAPS"><type>NULL</type></link> if - not found. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-stpcpy" role="function"> -<title>enca_stpcpy ()</title> -<indexterm zone="enca-stpcpy"><primary sortas="enca_stpcpy">enca_stpcpy</primary></indexterm><programlisting><link linkend="char">char</link>* enca_stpcpy (<link linkend="char">char</link> *dest, - const <link linkend="char">char</link> *src);</programlisting> -<para> -Appends a string to the end of another strings, returning pointer to -the terminating zero byte. -</para> -<para> -Will be defined as <link linkend="stpcpy"><function>stpcpy()</function></link> when system provides it. -</para> -<para> -Caller is responisble for providing <parameter>dest</parameter> long enough to hold the result.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>dest</parameter> :</term> -<listitem><simpara> A string. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>src</parameter> :</term> -<listitem><simpara> A string to append. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Pointer to the terminating zero byte of resulting string. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-strconcat" role="function"> -<title>enca_strconcat ()</title> -<indexterm zone="enca-strconcat"><primary sortas="enca_strconcat">enca_strconcat</primary></indexterm><programlisting><link linkend="char">char</link>* enca_strconcat (const <link linkend="char">char</link> *str, - ...);</programlisting> -<para> -Concatenates arbitrary (but at least one) number of strings.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>str</parameter> :</term> -<listitem><simpara> A string. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>...</parameter> :</term> -<listitem><simpara> A <link linkend="NULL-terminated"><type>NULL-terminated</type></link> list of string to append. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> All the strings concatenated together. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-strappend" role="function"> -<title>enca_strappend ()</title> -<indexterm zone="enca-strappend"><primary sortas="enca_strappend">enca_strappend</primary></indexterm><programlisting><link linkend="char">char</link>* enca_strappend (<link linkend="char">char</link> *str, - ...);</programlisting> -<para> -Appends arbitrary number of strings to a string. -</para> -<para> -The string <parameter>str</parameter> is destroyed (reallocated), the others are kept.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>str</parameter> :</term> -<listitem><simpara> A string. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>...</parameter> :</term> -<listitem><simpara> A <link linkend="NULL-terminated"><type>NULL-terminated</type></link> list of string to append. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> All the strings concatenated together. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-csname" role="macro"> -<title>enca_csname()</title> -<indexterm zone="enca-csname"><primary sortas="enca_csname">enca_csname</primary></indexterm><programlisting>#define enca_csname(cs) enca_charset_name((cs), ENCA_NAME_STYLE_ENCA) -</programlisting> -<para> -A shorthand for printing names with <link linkend="ENCA-NAME-STYLE-ENCA--CAPS"><type>ENCA_NAME_STYLE_ENCA</type></link>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>cs</parameter> :</term> -<listitem><simpara> A charset id. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-name-to-charset" role="function"> -<title>enca_name_to_charset ()</title> -<indexterm zone="enca-name-to-charset"><primary sortas="enca_name_to_charset">enca_name_to_charset</primary></indexterm><programlisting><link linkend="int">int</link> enca_name_to_charset (const <link linkend="char">char</link> *csname);</programlisting> -<para> -Transforms charset name to numeric charset id.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>csname</parameter> :</term> -<listitem><simpara> The charset name. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The charset id; <link linkend="ENCA-CS-UNKNOWN--CAPS"><type>ENCA_CS_UNKNOWN</type></link> when the name is not recognized. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-name-to-surface" role="function"> -<title>enca_name_to_surface ()</title> -<indexterm zone="enca-name-to-surface"><primary sortas="enca_name_to_surface">enca_name_to_surface</primary></indexterm><programlisting><link linkend="EncaSurface">EncaSurface</link> enca_name_to_surface (const <link linkend="char">char</link> *sname);</programlisting> -<para> -Transforms surface name to numeric surface id.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>sname</parameter> :</term> -<listitem><simpara> The surface name. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The surface id; <link linkend="ENCA-SURFACE-UNKNOWN--CAPS"><literal>ENCA_SURFACE_UNKNOWN</literal></link> when the name is not -recognized. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-language-init" role="function"> -<title>enca_language_init ()</title> -<indexterm zone="enca-language-init"><primary sortas="enca_language_init">enca_language_init</primary></indexterm><programlisting><link linkend="int">int</link> enca_language_init (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - const <link linkend="char">char</link> *langname);</programlisting> -<para> -Initializes analyser for language <parameter>langname</parameter>. -</para> -<para> -Assumes <parameter>analyser</parameter> is unitinialized, calling with an initialized <parameter>analyser</parameter> -leads to memory leak.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state to be initialized for this language. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>langname</parameter> :</term> -<listitem><simpara> Two-letter ISO-639 language code. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero on success, zero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-language-destroy" role="function"> -<title>enca_language_destroy ()</title> -<indexterm zone="enca-language-destroy"><primary sortas="enca_language_destroy">enca_language_destroy</primary></indexterm><programlisting><link linkend="void">void</link> enca_language_destroy (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Destroys the language part of analyser state <parameter>analyser</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state whose language part should be destroyed. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-get-charset-similarity-matrix" role="function"> -<title>enca_get_charset_similarity_matrix ()</title> -<indexterm zone="enca-get-charset-similarity-matrix"><primary sortas="enca_get_charset_similarity_matrix">enca_get_charset_similarity_matrix</primary></indexterm><programlisting><link linkend="double">double</link>* enca_get_charset_similarity_matrix (const <link linkend="EncaLanguageInfo">EncaLanguageInfo</link> *lang);</programlisting> -<para> -Computes character weight similarity matrix for language <parameter>lang</parameter>. -</para> -<para> -sim[i,j] is normalized to sim[i,i] thus: -- a row i contains ,probabilities` different languages will look like the - i-th one -- a column i contains ,probabilities` the i-th language will look like - the other languages. -</para> -<para> -For all practical applications, the higher one of sim[i,j] and sim[j,i] -is important. -</para> -<para> -Note: this is not used anywhere, only by simtable.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>lang</parameter> :</term> -<listitem><simpara> A language. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The matrix, its size is determined by <parameter>lang</parameter>->ncharsets; <link linkend="NULL--CAPS"><type>NULL</type></link> - for language with no charsets. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-charsets-subset-identical" role="function"> -<title>enca_charsets_subset_identical ()</title> -<indexterm zone="enca-charsets-subset-identical"><primary sortas="enca_charsets_subset_identical">enca_charsets_subset_identical</primary></indexterm><programlisting><link linkend="int">int</link> enca_charsets_subset_identical (<link linkend="int">int</link> charset1, - <link linkend="int">int</link> charset2, - const <link linkend="size-t">size_t</link> *counts);</programlisting> -<para> -Checks whether all characters with nonzero count have the same meaning -in both charsets. -</para> -<para> -In other words, it checks whether conversion of sample containing only -these characters from <parameter>charset1</parameter> to <parameter>charset2</parameter> would be identity.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>charset1</parameter> :</term> -<listitem><simpara> A charset. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>charset2</parameter> :</term> -<listitem><simpara> Another charset. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>counts</parameter> :</term> -<listitem><simpara> An array of size 0x100 containing character counts. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero if charsets are identical on the subset, zero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-filter-boxdraw" role="function"> -<title>enca_filter_boxdraw ()</title> -<indexterm zone="enca-filter-boxdraw"><primary sortas="enca_filter_boxdraw">enca_filter_boxdraw</primary></indexterm><programlisting><link linkend="size-t">size_t</link> enca_filter_boxdraw (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - unsigned <link linkend="char">char</link> fill_char);</programlisting> -<para> -Runs boxdrawing characters filter on <parameter>buffer</parameter> for each charset in <parameter>language</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser whose charsets should be considered for filtration. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>fill_char</parameter> :</term> -<listitem><simpara> Replacement character for filtered bytes. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Number of characters filtered out. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-language-hook-ncs" role="function"> -<title>enca_language_hook_ncs ()</title> -<indexterm zone="enca-language-hook-ncs"><primary sortas="enca_language_hook_ncs">enca_language_hook_ncs</primary></indexterm><programlisting><link linkend="int">int</link> enca_language_hook_ncs (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - <link linkend="size-t">size_t</link> ncs, - <link linkend="EncaLanguageHookData1CS">EncaLanguageHookData1CS</link> *hookdata);</programlisting> -<para> -Decide between two charsets differing only in a few characters. -</para> -<para> -If the two most probable charsets correspond to <parameter>hookdata</parameter> charsets, -give the characters they differ half the weight of all other characters -together, thus allowing to decide between the two very similar charsets. -</para> -<para> -It also recomputes <parameter>order</parameter> when something changes.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser whose charset ratings are to be modified. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>ncs</parameter> :</term> -<listitem><simpara> The number of charsets. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>hookdata</parameter> :</term> -<listitem><simpara> What characters of which charsets should be given the extra - weight. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when <parameter>ratings</parameter> were actually modified, nonzero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-language-hook-eol" role="function"> -<title>enca_language_hook_eol ()</title> -<indexterm zone="enca-language-hook-eol"><primary sortas="enca_language_hook_eol">enca_language_hook_eol</primary></indexterm><programlisting><link linkend="int">int</link> enca_language_hook_eol (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser, - <link linkend="size-t">size_t</link> ncs, - <link linkend="EncaLanguageHookDataEOL">EncaLanguageHookDataEOL</link> *hookdata);</programlisting> -<para> -Decide between two charsets differing only in EOL type or other surface. -</para> -<para> -The (surface mask, charset) pairs are scanned in order. If a matching -surface is found, ratings of all other charsets in the list are zeroed. -So you can place a surface mask of all 1s at the end to match when nothing -else matches. -</para> -<para> -All the charsets have to have the same rating, or nothing happens. -</para> -<para> -It also recomputes <parameter>order</parameter> when something changes.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser whose charset ratings are to be modified. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>ncs</parameter> :</term> -<listitem><simpara> The number of charsets. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>hookdata</parameter> :</term> -<listitem><simpara> What characters of which charsets should be decided with based - on the EOL type. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when <parameter>ratings</parameter> were actually modified, nonzero otherwise. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-guess-init" role="function"> -<title>enca_guess_init ()</title> -<indexterm zone="enca-guess-init"><primary sortas="enca_guess_init">enca_guess_init</primary></indexterm><programlisting><link linkend="void">void</link> enca_guess_init (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Allocates and initializes analyser state, sets options to defaults. -</para> -<para> -Assumes <parameter>analyser</parameter> is unitinialized, calling with an initialized <parameter>analyser</parameter> -leads to memory leak, but <parameter>analyser</parameter>->lang must be already initialized.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser to initialize. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-guess-destroy" role="function"> -<title>enca_guess_destroy ()</title> -<indexterm zone="enca-guess-destroy"><primary sortas="enca_guess_destroy">enca_guess_destroy</primary></indexterm><programlisting><link linkend="void">void</link> enca_guess_destroy (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Frees memory owned by analyser state.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyser to destroy. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-eol-surface" role="function"> -<title>enca_eol_surface ()</title> -<indexterm zone="enca-eol-surface"><primary sortas="enca_eol_surface">enca_eol_surface</primary></indexterm><programlisting><link linkend="EncaSurface">EncaSurface</link> enca_eol_surface (unsigned <link linkend="char">char</link> *buffer, - <link linkend="size-t">size_t</link> size, - const <link linkend="size-t">size_t</link> *counts);</programlisting> -<para> -Find EOL type of sample in <parameter>buffer</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>buffer</parameter> :</term> -<listitem><simpara> A buffer whose EOL type is to be detected. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>size</parameter> :</term> -<listitem><simpara> Size of <parameter>buffer</parameter>. -</simpara></listitem></varlistentry> -<varlistentry><term><parameter>counts</parameter> :</term> -<listitem><simpara> Character counts. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> The EOL surface flags. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-find-max-sec" role="function"> -<title>enca_find_max_sec ()</title> -<indexterm zone="enca-find-max-sec"><primary sortas="enca_find_max_sec">enca_find_max_sec</primary></indexterm><programlisting><link linkend="void">void</link> enca_find_max_sec (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Updates <parameter>analyser</parameter>->order according to charset <parameter>ratings</parameter>. -</para> -<para> -XXX: This should be stable sort. The ordering is defined by -data/<lang>/<lang>.h header file which is in turn defined by odering in -the appropriate script (doit.sh). Silly. -</para> -<para> -Must not be called with <parameter>analyser</parameter> with no regular charsets.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> An analyser. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="ENCA-LANGUAGE-BE--CAPS" role="variable"> -<title>ENCA_LANGUAGE_BE</title> -<indexterm zone="ENCA-LANGUAGE-BE--CAPS"><primary sortas="ENCA_LANGUAGE_BE">ENCA_LANGUAGE_BE</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_BE; -</programlisting> -<para> -Belarussian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-BG--CAPS" role="variable"> -<title>ENCA_LANGUAGE_BG</title> -<indexterm zone="ENCA-LANGUAGE-BG--CAPS"><primary sortas="ENCA_LANGUAGE_BG">ENCA_LANGUAGE_BG</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_BG; -</programlisting> -<para> -Bulgarian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-CS--CAPS" role="variable"> -<title>ENCA_LANGUAGE_CS</title> -<indexterm zone="ENCA-LANGUAGE-CS--CAPS"><primary sortas="ENCA_LANGUAGE_CS">ENCA_LANGUAGE_CS</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_CS; -</programlisting> -<para> -Czech language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-ET--CAPS" role="variable"> -<title>ENCA_LANGUAGE_ET</title> -<indexterm zone="ENCA-LANGUAGE-ET--CAPS"><primary sortas="ENCA_LANGUAGE_ET">ENCA_LANGUAGE_ET</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_ET; -</programlisting> -<para> -Estonian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-HR--CAPS" role="variable"> -<title>ENCA_LANGUAGE_HR</title> -<indexterm zone="ENCA-LANGUAGE-HR--CAPS"><primary sortas="ENCA_LANGUAGE_HR">ENCA_LANGUAGE_HR</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_HR; -</programlisting> -<para> -Croatian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-HU--CAPS" role="variable"> -<title>ENCA_LANGUAGE_HU</title> -<indexterm zone="ENCA-LANGUAGE-HU--CAPS"><primary sortas="ENCA_LANGUAGE_HU">ENCA_LANGUAGE_HU</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_HU; -</programlisting> -<para> -Hungarian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-LT--CAPS" role="variable"> -<title>ENCA_LANGUAGE_LT</title> -<indexterm zone="ENCA-LANGUAGE-LT--CAPS"><primary sortas="ENCA_LANGUAGE_LT">ENCA_LANGUAGE_LT</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_LT; -</programlisting> -<para> -Lithuanian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-LV--CAPS" role="variable"> -<title>ENCA_LANGUAGE_LV</title> -<indexterm zone="ENCA-LANGUAGE-LV--CAPS"><primary sortas="ENCA_LANGUAGE_LV">ENCA_LANGUAGE_LV</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_LV; -</programlisting> -<para> -Latvian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-PL--CAPS" role="variable"> -<title>ENCA_LANGUAGE_PL</title> -<indexterm zone="ENCA-LANGUAGE-PL--CAPS"><primary sortas="ENCA_LANGUAGE_PL">ENCA_LANGUAGE_PL</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_PL; -</programlisting> -<para> -Polish language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-RU--CAPS" role="variable"> -<title>ENCA_LANGUAGE_RU</title> -<indexterm zone="ENCA-LANGUAGE-RU--CAPS"><primary sortas="ENCA_LANGUAGE_RU">ENCA_LANGUAGE_RU</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_RU; -</programlisting> -<para> -Russian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-SK--CAPS" role="variable"> -<title>ENCA_LANGUAGE_SK</title> -<indexterm zone="ENCA-LANGUAGE-SK--CAPS"><primary sortas="ENCA_LANGUAGE_SK">ENCA_LANGUAGE_SK</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_SK; -</programlisting> -<para> -Slovak language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-SL--CAPS" role="variable"> -<title>ENCA_LANGUAGE_SL</title> -<indexterm zone="ENCA-LANGUAGE-SL--CAPS"><primary sortas="ENCA_LANGUAGE_SL">ENCA_LANGUAGE_SL</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_SL; -</programlisting> -<para> -Slovene language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="ENCA-LANGUAGE-UK--CAPS" role="variable"> -<title>ENCA_LANGUAGE_UK</title> -<indexterm zone="ENCA-LANGUAGE-UK--CAPS"><primary sortas="ENCA_LANGUAGE_UK">ENCA_LANGUAGE_UK</primary></indexterm><programlisting>extern const EncaLanguageInfo ENCA_LANGUAGE_UK; -</programlisting> -<para> -Ukrainian language. -</para> -<para> -Everything the world out there needs to know about this language.</para> -<para> -</para></refsect2> -<refsect2 id="enca-double-utf8-init" role="function"> -<title>enca_double_utf8_init ()</title> -<indexterm zone="enca-double-utf8-init"><primary sortas="enca_double_utf8_init">enca_double_utf8_init</primary></indexterm><programlisting><link linkend="void">void</link> enca_double_utf8_init (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Initializes double-UTF-8 check. -</para> -<para> -In fact it initializes the fields to <link linkend="NULL--CAPS"><type>NULL</type></link>'s, they are actually initialized -only when needed.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state to be initialized. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-double-utf8-destroy" role="function"> -<title>enca_double_utf8_destroy ()</title> -<indexterm zone="enca-double-utf8-destroy"><primary sortas="enca_double_utf8_destroy">enca_double_utf8_destroy</primary></indexterm><programlisting><link linkend="void">void</link> enca_double_utf8_destroy (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Destroys the double-UTF-8 check part of analyser state <parameter>analyser</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state whose double-UTF-8 check part should be destroyed. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-pair-init" role="function"> -<title>enca_pair_init ()</title> -<indexterm zone="enca-pair-init"><primary sortas="enca_pair_init">enca_pair_init</primary></indexterm><programlisting><link linkend="void">void</link> enca_pair_init (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Initializes pair statistics data. -</para> -<para> -In fact it just sets everything to <link linkend="NULL--CAPS"><type>NULL</type></link>, to be initialized when needed.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state to be initialized. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-pair-destroy" role="function"> -<title>enca_pair_destroy ()</title> -<indexterm zone="enca-pair-destroy"><primary sortas="enca_pair_destroy">enca_pair_destroy</primary></indexterm><programlisting><link linkend="void">void</link> enca_pair_destroy (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Destroys the pair statistics part of analyser state <parameter>analyser</parameter>.</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analyzer state whose pair statistics part should be destroyed. -</simpara></listitem></varlistentry> -</variablelist></refsect2> -<refsect2 id="enca-pair-analyse" role="function"> -<title>enca_pair_analyse ()</title> -<indexterm zone="enca-pair-analyse"><primary sortas="enca_pair_analyse">enca_pair_analyse</primary></indexterm><programlisting><link linkend="int">int</link> enca_pair_analyse (<link linkend="EncaAnalyserState">EncaAnalyserState</link> *analyser);</programlisting> -<para> -Performs pair-frequency based analysis, provided that the language supports -it (does nothing otherwise).</para> -<para> -</para><variablelist role="params"> -<varlistentry><term><parameter>analyser</parameter> :</term> -<listitem><simpara> Analysed containing the sample for pair frequency analysis. -</simpara></listitem></varlistentry> -<varlistentry><term><emphasis>Returns</emphasis> :</term><listitem><simpara> Nonzero when the character set was succesfully determined, - <parameter>analyser</parameter>-><parameter>result</parameter>.<parameter>charset</parameter> is then directly modified. -</simpara></listitem></varlistentry> -</variablelist></refsect2> - -</refsect1> - - - - -</refentry> diff --git a/lib/enca/devel-docs/xml/libenca-doc.bottom b/lib/enca/devel-docs/xml/libenca-doc.bottom deleted file mode 100644 index 27385094b8..0000000000 --- a/lib/enca/devel-docs/xml/libenca-doc.bottom +++ /dev/null @@ -1,5 +0,0 @@ - <xi:include href="xml/enums.xml"/> - <xi:include href="xml/analyser.xml"/> - <xi:include href="xml/encodings.xml"/> - <xi:include href="xml/auxiliary.xml"/> - <xi:include href="xml/internal.xml"/> diff --git a/lib/enca/devel-docs/xml/libenca-doc.top b/lib/enca/devel-docs/xml/libenca-doc.top deleted file mode 100644 index e69de29bb2..0000000000 --- a/lib/enca/devel-docs/xml/libenca-doc.top +++ /dev/null diff --git a/lib/enca/devel-docs/xml/object_index.sgml b/lib/enca/devel-docs/xml/object_index.sgml deleted file mode 100644 index 229837e7ef..0000000000 --- a/lib/enca/devel-docs/xml/object_index.sgml +++ /dev/null @@ -1,11 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE informaltable PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<informaltable pgwide="1" frame="none"> -<tgroup cols="3"> -<colspec colwidth="1*"/> -<colspec colwidth="1*"/> -<colspec colwidth="1*"/> -<tbody> -<row><entry> </entry></row> -</tbody></tgroup></informaltable> diff --git a/lib/enca/devel-docs/xml/tree_index.sgml b/lib/enca/devel-docs/xml/tree_index.sgml deleted file mode 100644 index 6f979c8836..0000000000 --- a/lib/enca/devel-docs/xml/tree_index.sgml +++ /dev/null @@ -1,5 +0,0 @@ -<?xml version="1.0"?> -<!DOCTYPE screen PUBLIC "-//OASIS//DTD DocBook XML V4.1.2//EN" - "http://www.oasis-open.org/docbook/xml/4.1.2/docbookx.dtd"> -<screen> -</screen> |